Influence of Robotic Interventions on Gait Improvement in Children with Cerebral Palsy (2025)

For individuals with cerebral palsy (CP), walking ability is of critical importance, as highlighted by the focus on gross motor function within their primary outcome measure, the Gross Motor Function Classification System (GMFCS). This focus underscores the intricate connection between walking ability and participation, activity, and physical function. Despite extensive and prolonged therapeutic interventions, rehabilitation efforts often fail to produce significant improvements in walking ability for individuals with CP. Recently, robotic-assisted gait training (RAGT) has emerged as a promising therapeutic modality for enhancing walking capabilities in this population. RAGT offers the potential for personalized interventions by adjusting parameters such as assistance level, resistance, and body weight support to cater to the specific needs of individuals with CP. Nevertheless, the evidence supporting the efficacy of RAGT remains limited. This chapter comprehensively reviews the challenges associated with walking in individuals with CP, explores the potential benefits and various forms of RAGT, and discusses future research directions.

Author Information

Show +

• Lihua Jin

  • Department of Rehabilitation Medicine, Second Affiliated Hospital of Kunming Medical University, Kunming,Yunnan, China

• Caixia Zhao

  • Department of Pediatrics, Yunnan University of Traditional Chinese Medicine, Kunming,Yunnan, China

• Binjing Dou

  • Department of Pediatrics, Yunnan University of Traditional Chinese Medicine, Kunming,Yunnan, China

• Juchuan Dong

  • Department of Rehabilitation Medicine, Second Affiliated Hospital of Kunming Medical University, Kunming,Yunnan, China

• Ping He*

  • Department of Pediatrics, First Affiliated Hospital of Yunnan University of Traditional Chinese Medicine, Kunming,Yunnan, China

*Address all correspondence to: hepingx8@sina.com

1. Introduction

Cerebral palsy (CP), caused by brain damage prenatally, perinatally, or postnatally, is the most common cause of childhood movement and posture-related disability. Motor impairments manifest in approximately 80% of children with CP, resulting in secondary complications such as hip pain, dislocation, balance issues, hand dysfunction, and equinus deformity [1]. The brain damage associated with CP can result in positive symptoms, including muscle spasms and hypertonia, and/or negative symptoms, such as muscle weakness, loss of motor control, muscle wasting, and impaired balance reactions. The predominant motor disorders associated with CP include spasticity and abnormalities in muscle tone, which are commonly associated with challenges in coordination, strength, and selective motor control. These conditions can result in spasticity-induced deformities of the bones and joints, as well as pain and functional impairment [2]. Muscle injuries arising from spasms, weakness, abnormal tone, or pain can further diminish joint range of motion and selective motor control. Consequently, difficulties regulating the flexion angles of the lower limb joints, specifically the hip, knee, and ankle, impede the ambulatory capabilities of affected children, thereby impacting on daily activities. Gait abnormalities, including crouch gait, toe walking, and scissoring, are frequently observed in children with CP and are often linked to diminished participation in daily activities and social interactions, underscoring the importance of maintaining and enhancing walking ability throughout all stages of CP management.

Abnormal gait associated with CP can present in various forms, which are predominantly categorized as pathological gait patterns. For example, individuals with spastic CP display a reduction in the muscle volume of the gastrocnemius and tibialis anterior, accompanied by an increased echo intensity, indicative of abnormal ankle gait characteristics [3]. Patients with CP also exhibit a reduced number of muscle synergies during gait, indicating the adoption of a simplified control strategy during ambulation, potentially linked to diminished neuromuscular control complexity. Among younger patients with CP, the most prevalent gait abnormality is true equinus gait [4]. At this developmental stage, the patient’s body weight is relatively low, and the muscle strength of the hip and knee joints remains robust; however, plantar flexion is caused by the shortening of the gastrocnemius muscle group. This abnormal gait is characterized by forefoot contact during the stance phase, with normal flexion and extension of the hip and knee joints. As patients with CP age, their body weight increases, concurrent with the shortening or weakening of the proximal muscles. This results in significant flexion of the hip and knee joints during the single-stance phase, resulting in a gait pattern that resembles a jump. As this proximal muscle weakness progresses, supporting the body weight becomes increasingly complicated, resulting in a more pronounced flexion of the hip and knee joints, leading to the development of a distinct equinus gait. Eventually, when patients with CP experience dorsiflexion of the foot and ankle, along with severe flexion of the hip and knee joints, the gait transitions into a crouch gait. Crouch gait, the ultimate developmental stage of gait dysfunction in these patients, is also the prevalent gait abnormality observed in children with CP [5]. These disparate gait patterns all develop as compensatory mechanisms for insufficient muscle strength to support body weight. However, all are unsustainable and tend to deteriorate as the patient grows and gains weight, ultimately resulting in a loss of motor function. Consequently, it is imperative to devise strategies aimed at maintaining long-term gait function to enhance the rehabilitation outcomes of patients with CP.

In this chapter, we discuss the use of robotic-assisted gait training (RAGT) as the potential intervention in the improvement of locomotor function in patients with CP.

2. Gait enhancement strategies in children with CP

Due to the heterogeneous nature of CP, patients with CP exhibit a diverse array of associated challenges and motor impairments that necessitate evaluation and intervention by a multidisciplinary team of specialists. Enhancing gait is a critical therapeutic objective in CP management; however, conventional gait rehabilitation methods are often unsustainable for patients. While early initiation of rehabilitation, particularly through a combination of physical and occupational therapy, may enhance outcomes [6], conventional rehabilitation approaches have demonstrated limited efficacy in the treatment of CP. This limitation is especially pertinent given that rehabilitation largely depends on the brain’s neuroplasticity, which tends to be less responsive to interventions in adult individuals with CP [7]. Notably, adults with CP may derive cardiovascular benefits from physical therapy interventions, such as aerobic exercise and resistance training; however, significant improvements in walking ability are unlikely to be achieved [8]. Other potential strategies include surgical interventions, whose effects are often transient, losing efficacy as children grow and develop, and manual assistance, which is limited by its demanding, intensive, and continuous nature. Robotic technology is an innovative rehabilitation approach that employs computer-controlled systems to facilitate motor learning and cortical reorganization, thereby enhancing limb function.

3. An introduction to RAGT

Prior studies have shown that gait rehabilitation incorporating functional movements has greater efficacy than training employing nonfunctional movement patterns. RAGT offers precise and intensive task-specific training, while simultaneously collecting data from multiple sensors to accurately capture patient information, such as limb kinematics, kinetics, electromyography (EMG) pattern, electroencephalography (EEG) patterns, and energy expenditure, during practice. Thus, the implementation of robotic-assisted therapy has the potential to alleviate skilled therapists from the physical demands associated with direct hands-on therapy. Furthermore, RAGT offers repetitive, continuous, graded, and task-specific training aimed at improving gait function by establishing conditions conducive to motor learning principles, such as intensity, repetition, task specificity, and engagement. As such, this rehabilitation strategy facilitates neuroplastic changes and enhances non-locomotor recovery in children with CP experiencing movement disorders of a central origin.

Repetitive RAGT fosters functional network reorganization within the sensorimotor cortex and stimulates neuroplasticity [9] while enhancing functional connectivity between the frontal and parietal regions [10]. The mechanisms by which RAGT influences the brain are intricate, involving multiple neural pathways and cortical regions. Early initiation of RAGT following brain injury can markedly expedite the bilateral reorganization of motor-related brain areas [11]. Studies employing functional near-infrared spectroscopy (fNIRS) to monitor changes in the cerebral cortex during RAGT have revealed that prompt RAGT intervention can significantly enhance cortical activation associated with motor control; indeed, this activation extends beyond the motor areas to encompass various functional brain regions [12]. Furthermore, RAGT has the potential to modulate the overall efficiency of brain networks, facilitating the restoration of normal brain network organization, and thereby promoting functional recovery [13]. Together, these findings indicate that RAGT holds significant potential for improving motor function and quality of life among children with CP.

RAGT integrates exoskeleton technology with information control systems to aid in human movement through precise mechanical devices. This approach necessitates interdisciplinary expertise, encompassing biomechanics, robotics, information science, and artificial intelligence. Typically, these robotic systems are outfitted with sensors capable of performing real-time, high-precision assessments of patients’ rehabilitation status, including metrics such as joint range of motion, force output, walking speed, and step length. Furthermore, these systems can autonomously adapt to each patient, based on individual data, thereby allowing the personalization of rehabilitation programs [14].

4. Forms of RAGT

Functionally, RAGT can be categorized into two primary types: traction-type and exoskeleton-type. The defining characteristic of traction-type RAGT is its ability to facilitate the training of the hip, knee, and ankle joints by directing movement of the patient’s feet. The end-effector traction-type robot does not require alignment of its hardware system with the patient’s joint structure, thereby enhancing its user-friendliness. Consequently, this type of rehabilitation robot is recognized for its high safety standards during training sessions. In contrast, exoskeleton-type RAGT is characterized by a more intricate structure with advanced functionalities and capabilities. Compared to traction-type robots, exoskeleton-type can interpret each patient’s movement intentions, simulating and controlling multiple joints based on normal gait patterns, thereby offering personalized support. Consequently, exoskeleton-type RAGT is regarded as an emerging research direction and developmental trend within the rehabilitation domain.

RAGT integrates bilateral robotic orthoses, body weight support (BWS), and a treadmill. As a computerized system, it allows for the adjustment of BWS to ensure an upright posture and precise lower limb loading. The orthoses used in RAGT facilitate leg movements within the sagittal plane, following repeatable, predefined trajectories of the hip and knee joints. Simultaneously, footplates maintain passive ankle dorsiflexion, thereby assisting individuals with CP in performing ambulation. RAGT offers a controlled and safe therapeutic environment, enabling patients to participate in prolonged training sessions through numerous repetitions of steps while promoting repeatable, kinematically consistent symmetrical gait patterns [15].

5. Clinical utility of RAGT

The most commonly employed form of RAGT in clinical practice involves the use of tethered exoskeleton systems, which exert force through a rigid, articulated frame that facilitates the movement of the patient’s legs across one or more planes via a body weight support system [16], such as the Lokomat (Hocoma AG, Volketswil, Switzerland) system. Nevertheless, current supporting evidence remains insufficient to demonstrate that these systems yield superior therapeutic outcomes in terms of gross motor function and walking ability compared with conventional physiotherapy [17]. The primary limitation of these systems includes their reliance on predetermined fixed trajectories or timings, which restricts the patient’s active participation consequently diminishing muscle activity. Furthermore, when patients attempt to walk actively and resist the device, abnormal muscle activation patterns may develop [18]. In contrast, untethered exoskeletons, such as ReWalk (ReWalk Robotics Inc., Marlborough, MA, USA), Indego (Parker Hannifin Corp, Mayfield Heights, OH, USA), Hybrid Assistive Limb (HAL, Cyberdyne Inc., Tsukuba, Japan), and Ekso (Ekso Bionics, Richmond, CA, USA), are wearable systems comprising articulated suits that are self-powered and equipped with advanced control algorithms. These systems, which are used for gait assistance and rehabilitation, provide task-specific overground training to offer patients a more liberated and realistic walking experience.

6. Clinically available RAGT devices

Various RAGT devices offer stability and partial body weight support for pediatric patients engaged in overground walking training. By enabling the individualized control of each joint, these devices can operate in multiple modes, including position control, resistance, and zero-force control, thereby allowing movements to be tailored to each patient’s current capabilities and enhancing the system’s modularity [19], as exemplified by devices such as the CP Walker. The Innowalk Pro (IP) [20] is a standing apparatus designed to facilitate movement across different spatial dimensions. This device adjusts the child’s range of motion, frequency, and functionality during use to optimize therapeutic outcomes [21]. The Walkbot-K system includes both adult and pediatric exoskeleton RAGT models, with the pediatric version specifically designed for children with CP. This device can automatically adjust leg length to fit the child, thereby promoting functional improvement, while providing real-time assessments of muscle tone, muscle strength, and gait [22]. The pediatric knee exoskeleton (P.REX) is the second prototype of a tethered knee exoskeleton; the P.REX comprises an untethered device capable of delivering consistent knee extension torque throughout various movement phases [23]. The Honda Walking Assist (HWA) device, a wearable exoskeleton robot developed by Honda R&D Co., Ltd., was designed to facilitate bilateral hip flexion and extension during ambulation. Notably, the HWA targets a single joint without constraining the degrees of freedom of other joints, thereby permitting a high degree of movement and enhancing motor learning efficacy. HWA has been shown to significantly reduce energy expenditure in healthy young adults and can improve hip kinematic symmetry and step length in individuals with hemiparesis. Considering that the majority of children with CP exhibit bilateral neurological involvement and complicated lower limb symptoms, the HWA holds potential to aid these children in acquiring symmetrical gait patterns by supporting bilateral hip movements [20].

7. Clinical evidence supporting the utility of RAGT

Numerous studies investigating RAGT in populations with CP have employed systems using tethered exoskeletons, reporting consequent enhancements in gait speed, endurance, and gross motor function [24]. However, meta-analytical evaluations have deemed these findings statistically insignificant. A recent systematic review has highlighted that the evidence supporting the efficacy of RAGT for children with gait disorders, particularly CP, remains weak and inconsistent, with RAGT not demonstrating superior outcomes compared to traditional physical therapy [25]. Additional research has shown that RAGT does not significantly enhance standing ability or gait function in patients with CP classified as levels III to IV on the Gross Motor Function Classification System (GMFCS) [26]. Additionally, when motor function levels were not taken into account, randomized controlled trials assessing gait training interventions in patients with CP indicated that RAGT did not demonstrate superiority over conventional physical therapy [27]. Furthermore, studies have indicated that the harness-provided weight support and treadmill belt movement may render RAGT training more demanding and intense than overground walking [28]. Finally, the observed improvements in gait function may not translate effectively to overground walking, as treadmill training lacks task specificity for real-world environments.

Despite some skepticism, numerous studies have documented the beneficial effects of RAGT on gait in patients with CP. These benefits include improvements in joint range of motion [29, 30, 31], muscle tone [30], muscle strength [29], balance [26], and short-term and long-term gait parameters, particularly gait speed [32]. Additionally, enhancements in gross motor function [26], particularly in patients classified as GMFCS levels IV and V, gait kinematics [33, 34], reduced metabolic cost during walking [35], and improved motor performance and endurance [16] have been observed. RAGT facilitates task-oriented repetitive movements, muscle strengthening, and motor coordination, all of which positively influence energy efficiency, gait speed, and balance control [36].

The variability in RAGT outcomes between disparate studies may be attributed to differences in parameter settings, as RAGT combined with virtual reality achieved enhanced efficacy at improving BWS when body weight was reduced by 30% [37]. Additionally, the efficacy of RAGT may be influenced by the training mode employed. For example, the motor and constraint-induced movement therapy (CIMT) modes were more effective at enhancing balance and gait [38]. The intensity and frequency of RAGT are critical determinants of rehabilitation outcomes which may directly impact alterations in patients’ motor function. Tailoring the intensity of RAGT based on the GMFCS level presents a promising intervention strategy, as higher-intensity training has shown more substantial effects in individuals classified at GMFCS levels II and III. However, prior studies on RAGT have employed diverse training protocols and implemented differing procedures, complicating the comparison of results across studies. This heterogeneity in training protocols may have contributed to the limited clinical evidence supporting RAGT. Nevertheless, compared to other robotic interventions, such as patient-guided suspension systems and end-effector devices, RAGT has been shown to offer comprehensive control over leg joint angles and torques, making it the preferred robotic solution for training patients with severe motor impairments due to brain injury [39].

In addition to RAGT application as a standalone intervention, promising outcomes have been observed for the combination of RAGT with other therapeutic modalities. For example, the integration of RAGT with botulinum toxin A (BoNT-A) enhanced gross motor function measure scores in children with CP, although RAGT did not amplify the anti-spasticity effects of BoNT-A. Nevertheless, its adjunctive use resulted in significant improvements in motor skills and gait in this population [40]. The combination of RAGT with virtual reality (VR) has further been reported to increase patient engagement and enjoyment during training, thereby effectively enhancing gait in children with CP [37]. Furthermore, RAGT in conjunction with noninvasive brain stimulation (NIBS) has shown the potential to improve lower limb function in various neurological populations [41]. Specifically, the combination of RAGT with repetitive transcranial magnetic stimulation (rTMS) can modulate cortical motor inhibition, resulting in gait improvements [42]. Additionally, the use of RAGT alongside transcranial direct current stimulation (tDCS) enhances balance and functional performance in individuals with CP [43].

Numerous studies investigating RAGT in populations with CP have employed systems using tethered exoskeletons, reporting consequent enhancements in gait speed, endurance, and gross motor function [24]. However, meta-analytical evaluations have deemed these findings statistically insignificant. A recent systematic review has highlighted that the evidence supporting the efficacy of RAGT for children with gait disorders, particularly CP, remains weak and inconsistent, with RAGT not demonstrating superior outcomes compared to traditional physical therapy [25]. Additional research has shown that RAGT does not significantly enhance standing ability or gait function in patients with CP classified as levels III to IV on the Gross Motor Function Classification System (GMFCS) [26]. Additionally, when motor function levels were not taken into account, randomized controlled trials assessing gait training interventions in patients with CP indicated that RAGT did not demonstrate superiority over conventional physical therapy [27]. Furthermore, studies have indicated that the harness-provided weight support and treadmill belt movement may render RAGT training more demanding and intense than overground walking [28]. Finally, the observed improvements in gait function may not translate effectively to overground walking, as treadmill training lacks task specificity for real-world environments.

Despite some skepticism, the beneficial effects of RAGT on various bodily functions in patients with CP cannot be dismissed. These benefits include improvements in joint range of motion [29, 30, 31], muscle tone [30]. muscle strength [29], balance [26], and short-term and long-term gait parameters, particularly gait speed [32]. Additionally, enhancements in gross motor function [26], particularly in patients classified as GMFCS levels IV and V, gait kinematics [33, 34], reduced metabolic cost during walking [35], and improved motor performance and endurance [16] have been observed. RAGT facilitates task-oriented repetitive movements, muscle strengthening, and motor coordination, all of which positively influence energy efficiency, gait speed, and balance control [36]. Given that long-term motor functional impairments can result in psychological issues, such as anxiety and depression, which may further exacerbate physical motor function, the efficacy of interventions targeting mental health is critical in the rehabilitation process. RAGT has been shown to enhance psychological well-being in patients with neurological impairments and positively affect short-term depressive symptoms.

The variability in RAGT outcomes between disparate studies may be attributed to differences in parameter settings, as RAGT combined with virtual reality achieved enhanced efficacy at improving BWS when body weight was reduced by 30% [37]. Additionally, the efficacy of RAGT may be influenced by the training mode employed. For example, the motor and constraint-induced movement therapy (CIMT) modes were more effective at enhancing balance and gait [38]. The intensity and frequency of RAGT are critical determinants of rehabilitation outcomes, which may directly impact alterations in patients’ motor function. Tailoring the intensity of RAGT based on the GMFCS level presents a promising intervention strategy, as higher-intensity training has shown more substantial effects in individuals classified at GMFCS levels II and III. However, prior studies on RAGT have employed diverse training protocols and implemented differing procedures, complicating the comparison of results across studies. This heterogeneity in training protocols may have contributed to the limited clinical evidence supporting RAGT. Nevertheless, compared to other robotic interventions, such as patient-guided suspension systems and end-effector devices, RAGT has been shown to offer comprehensive control over leg joint angles and torques, making it the preferred robotic solution for training patients with severe motor impairments due to brain injury [39].

8. Integration of RAGT with additional interventional modalities

Within conventional rehabilitation frameworks, RAGT is frequently integrated with traditional physical therapy techniques, such as joint and muscle relaxation exercises, to improve joint flexibility. These exercises may be administered either prior to or following RAGT to enhance the therapeutic outcomes of both physical therapy and RAGT. This combined approach has been shown to consistently increase lower limb muscle strength, mitigate muscle atrophy, and enhance muscle endurance in patients. Additionally, the concurrent use of RAGT with pharmacological interventions is prevalent in clinical settings, particularly as patients with CP often require ongoing oral medications, including antiepileptics and muscle relaxants. The administration of these medications during RAGT sessions can enhance safety and improve patient adherence to RAGT movement protocols, thereby augmenting the efficacy of the training. In addition to the interventions outlined in Section 7 that can be integrated with RAGT, the RAGT apparatus itself, when equipped with training modules, offers potential benefits for patients with central nervous system injuries. For example, beyond the conventional linear walking patterns associated with traditional RAGT, the incorporation of motion capture devices for multi-directional tracking—including left, right, and bilateral multi-stage trajectories—can accommodate patients at various stages of injury recovery [44]. Furthermore, the integration of RAGT with acupuncture therapy may offer promise, potentially assisting patients with abnormal gait and walking difficulties [45]. With advancements in Transcutaneous Spinal Cord Stimulation (tSCS) technology, its combined application with RAGT has demonstrated improvements in gait-related parameters for patients with CP, such as enhanced upright posture and reduced crouching [46]. Nevertheless, the empirical support for these adjunctive treatment methods remains limited. Electromyography (EMG), when integrated with RAGT, emerges as a promising auxiliary instrument for assessing the origins of abnormal postures during gait interventions and for informing clinical decisions on targeted combined approaches [47]. However, the technical complexity of EMG poses significant challenges to its widespread application in clinical settings. Consequently, further investigation is warranted to identify more feasible enhancements for RAGT modules and its combined approaches to improve therapeutic efficacy.

9. Accessibility of RAGT

In recent years, RAGT has been increasingly investigated for its application in pediatric rehabilitation within several advanced nations, yielding significant outcomes that affirm its technical feasibility, clinical efficacy, and potential for promising development. The global exchange of medical technology has facilitated opportunities for the introduction of RAGT equipment in developing countries. International medical aid initiatives and charitable organizations may contribute to this effort by donating or funding the acquisition of RAGT equipment for healthcare facilities in these regions, thereby progressively improving local accessibility. Although certain developing countries have endeavored to independently design and adapt RAGT equipment, achieving therapeutic benefits for patients with cerebral palsy [48, 49], the high costs associated with RAGT equipment and its maintenance, coupled with the operational complexity and requirement for specialized personnel, constrain its widespread adoption in grassroots hospitals and less developed regions [50]. Furthermore, the unequal distribution of treatment resources, particularly for rehabilitation technologies such as RAGT that necessitate substantial financial and technical investment, exacerbates existing regional disparities [51]. One approach to mitigating this disparity involves developing specialized expertise to serve regional populations using shared resources.

10. Conclusion

In summary, although significant strides have been made in the development of sophisticated motor learning-driven controllers aimed at enhancing gait rehabilitation, we believe that exoskeleton technology remains a sufficiently expansive and fruitful area of research. Given the myriad benefits associated with improved walking ability in patients with CP, as well as the promising functional enhancements reported by various RAGT systems, further exploration in this domain is likely to yield additional advantages for ambulatory function [52]. However, in order to fully clarify the benefits of RAGT, further multicenter studies with large sample sizes and uniform methodologies should be conducted. In addition, further investigation is required to clarify the optimal timing and rehabilitation strategies to maximize gait rehabilitation in patients with CP. Furthermore, we propose that as artificial intelligence progresses, future research should advance toward more structured and standardized investigations to elucidate and expand the RAGT functionalities advantageous for CP patients.