Robotic Gait Training in an Adolescent with Idiopathic Transverse Myelitis: A Case Report ^†

Abstract

Transverse myelitis is a rare spinal cord condition that can cause severe motor, sensory, and autonomic dysfunction. This case report describes a 16-year-old male with incomplete paraplegia due to idiopathic transverse myelitis who underwent robotic-assisted gait training (RAGT) using the EKSO exoskeleton, integrated into an intensive rehabilitation programme. After one month, he showed significant improvements in gait speed, dynamic balance, effort tolerance, and trunk mobility. RAGT promoted better weight distribution and reduced compensatory patterns during ambulation. The intervention proved safe and clinically beneficial, highlighting the potential of robotic technologies as effective adjuncts in paediatric spinal cord injury rehabilitation.

1. Introduction

Transverse myelitis (TM) is a rare, rapidly progressing inflammatory disorder of the spinal cord, characterised by demyelination and resulting neurological dysfunction below the level of the lesion. It typically presents with motor weakness, sensory deficits at a defined level, and autonomic disturbances [1,2]. In paediatric populations, TM has an estimated incidence of 1 to 3 cases per million children per year, most commonly affecting those under the age of five and adolescents [2].

Walking is a key milestone in TM recovery, particularly in children, as it supports independence. Gait impairments depend on the lesion’s level and severity, which affect motor control, strength, and tone in the trunk and lower limbs (LLs), which reduce participation in daily activities at home, school, and in the community [3,4,5].

Robotic-Assisted Gait Training (RAGT) is a promising strategy for locomotor recovery, leveraging neuroplasticity—the central nervous system’s ability to reorganise structurally and functionally in response to repetitive, task-specific input. Children, with their heightened neuroplastic potential, may respond well to such interventions; however, current evidence remains limited and heterogeneous [3,5]. There are various types of RAGT depending on the level of support provided. Robotic exoskeletons were designed to assist lower limb (LL) movement and support sit-to-stand transitions and walking [4,5], supporting gait rehabilitation by enhancing propulsion, reducing body weight load, and improving proprioceptive input. Upright positioning and bilateral LL loading are critical for stimulating proprioceptors and activating spinal pathways. This sensory input—reinforced through progressive weight-bearing and hip extension—promotes sensorimotor integration. By minimising compensatory movements, exoskeleton-assisted training enhances postural control and supports upper limb balance. Additionally, it facilitates retention of functional walking patterns [2]. Furthermore, in individuals with SCI patients, has shown greater benefits than conventional therapy, improving balance, effort tolerance, strength, spasticity, and respiratory function [4,6]. It is recommended that RAGT sessions last approximately 25 min [2].

This case study aims to demonstrate, over the course of a one-month rehabilitation programme, the effectiveness of integrating a robotic exoskeleton into the physiotherapy protocol to improve gait quality and spatiotemporal parameters in paediatric patients with incomplete SCI.

2. Materials and Methods

2.1. Participant

The participant was a 16-year-old male with D8 SCI, diagnosed with idiopathic TM involving the conus medullaris, classified according to the American Spinal Injury Association Impairment Scale (AIS) as AIS B paraplegia, with a neurological level of T11. He was 4 months post-injury at the time of enrolment, 1.71 m tall, and weighed 53.6 kg. Data collection was carried out with the informed consent of the participant and his parents, including explicit authorization for the collection and disclosure of clinical information for scientific research purposes and subsequent publication.

Prior to initiating RAGT, the participant underwent a detailed clinical assessment. No limitations were observed in joint range of motion or upper limb muscle strength. However, muscular weakness was evident in the trunk and the right LL, along with a complete absence of active movement in the left LL. Sensory function was preserved above the level of T11 in the spinal cord injury and reduced below, with greater impairment observed in the left LL. From a functional standpoint, was independent in bed mobility but required assistance for positional changes, relying heavily on compensatory overuse of the right LL. The participant was able to independently assume and maintain a seated position with effective postural balance. Upright posture was achieved with assistance and the use of a knee–ankle–foot orthosis (KAFO) on the left LL, compensating for the absence of active motor control. In standing, the participant required additional support to maintain posture or perform tasks due to limited postural control. No active weight transfer was observed onto the left LL, resulting in overuse of the right hemibody during functional tasks. This was accompanied by fixation and compensatory movement patterns on the right side. Even with external support, the participant was unable to achieve single-limb stance on the left side. Ambulation was possible over short distances on flat surfaces, using two elbow crutches in combination with a KAFO. The gait pattern was markedly asymmetrical, with overuse of the right hemibody. The right stance phase was prolonged, with excessive weight transfer compensating for the difficulty in initiating left step advancement. Conversely, the left swing phase was achieved through compensatory patterns, including ipsilateral pelvic hike, quadratus lumborum activation, and elevation of the upper right trunk to enable step circumduction. The left stance phase was nearly absent, as the participant was unable to advance the hip or initiate extension, which impeded effective weight transfer and consequently led to a shortened right swing phase. Quantitative gait assessments—10-Metre Walk Test (10MWT) for gait speed, the Timed Up and Go (TUG) test for dynamic balance, the 6-Minute Walk Test (6MWT) and the Rate of Perceived Exertion (RPE) for evaluating effort tolerance and perceived exertion—along with the Berg Balance Scale (BBS) for overall balance—are presented in

Table 1. Pre- and post-intervention spatiotemporal balance performance and gait parameters; T1: initial evaluation, T2: final evaluation, BBS: Berg Balance Scale, 10MWT: 10-Metre Walk Test, TUG: Timed Up and Go, 6MWT: 6 Minute Walk Test, RPE: Rate of Perceived Exertion.

	BBS	10MWT		TUG		6MWT
		Time (s)	Velocity (m/s)	Time (s)	Velocity (m/s)	Distance (m)	Velocity (m/s)	Initial Perceived Exertion (RPE)	Final Perceived Exertion (RPE)
T1	14/56	22	0.45	22	0.27	163	0.45	0/10	4/10
T2	25/56	18	0.56	17	0.35	205	0.57	0/10	1/10

.

2.2. Rehabilitation Protocol

The participant underwent an intensive multidisciplinary rehabilitation programme in a paediatric inpatient unit, with goals integrated into a personalised plan based on recovery stage and functional needs. Physiotherapy was provided twice daily on weekdays, totalling three hours per day, by the same therapist to ensure continuity.

The physiotherapy intervention aimed to enhance functional capacity and autonomy, in accordance with current clinical guidelines for incomplete SCI. The approach was grounded in principles of neuroplasticity-based rehabilitation, incorporating meaningful, task-oriented activities. Included manual therapy and positioning, strength and endurance training, balance and mobility exercises, and functional activity training [7]. After one month of inpatient rehabilitation, RAGT sessions were introduced at a frequency of two to three times per week. Each session lasted 30 min and included the setup of the exoskeleton, tailored to the participant’s physical characteristics and aligned with session goals. The intervention focused on overground gait training on a flat surface, and all sessions were conducted by two physiotherapists with specialised training in RAGT.

3. Results

Four months post-injury, Ekso was introduced to support sensorimotor integration, postural stability, and symmetrical weight distribution during gait, with a focus on improving left lower limb (LL) weight transfer, while using a KAFO. Over one month, patient completed 11 RAGT sessions (mean effective walking time: 15.64 ± 4.7 min), showing functional improvements following the combined physiotherapy and RAGT programme.

Improvements in biomechanical gait alignment were evident through clinical observation, including enhanced weight-bearing capacity on the more affected lower limb and improved sensorimotor integration. This progress enabled the patient to achieve a supported single-limb stance on the left lower limb. Although this did not yet permit the execution of specific functional tasks, it represents a positive prognostic indicator for future recovery. The structured and pre-programmed nature of RAGT facilitated safe and effective weight-shifting, allowing for more balanced lower limb loading and improved interlimb coordination to be observed. In addition, it contributed to an overall improvement in movement patterns by reducing the need for compensatory strategies. Prior to the intervention, the patient demonstrated significant difficulty bearing weight on the left LL, resulting in compensatory trunk movements to maintain gait. Post-intervention, a reduction in these compensatory strategies was observed, along with improved trunk mobility and greater dissociation between the shoulder and pelvic girdles. Improvements in gait rhythm were also noted, contributing to a more automated and fluent walking pattern.

In addition to qualitative changes, quantitative gait parameters demonstrated functionally relevant improvements, as shown in Table 1. Gait speed increased, as demonstrated by the 10MWT result, surpassing the Minimal Clinically Important Difference (MCID) [8], enhancing day-to-day functional capacity and walking efficiency. Effort tolerance improved, as evidenced by the 6MWT results, reflecting greater functional endurance during daily activities. This enhancement exceeded the MCID [9], indicating a clinically meaningful gain in walking capacity. Furthermore, the perceived exertion during the test also decreased, suggesting an increased tolerance to physical effort. Dynamic balance also improved, particularly during gait as showed in TUG results, contributing to greater safety and a reduced risk of falls. A significant improvement in overall functional balance was also observed with the BBS score, which reflects not only improved gait stability but also enhanced performance in other balance-related activities. No episodes of excessive fatigue or adverse events were reported during the RAGT sessions.

4. Discussion

While robotic exoskeletons are primarily standardised to improve gait parameters [2], this case illustrates their potential to address broader motor goals related to overall gait pattern quality—including enhanced sensorimotor integration of the LLs, more effective weight transfer, and reduced reliance on compensatory strategies, as observed in the patient’s gait. The patient’s increased effort tolerance, along with improvements in gait velocity and dynamic balance (Table 1), further support this broader therapeutic value.

In addition to the qualitative observations, quantitative outcome measures also demonstrated significant progress. Gait speed increased, dynamic and functional balance improved, and effort tolerance rose across sessions. Importantly, some of the gains observed in walking assessments exceeded the MCID thresholds, reinforcing the clinical relevance of the intervention. The patient’s BBS score improved, suggesting enhanced postural control, even though orthotic dependency and absence of movement in the left LL limited maximum scoring potential. Although the final BBS score did not reach the 40-point threshold associated with reduced fall risk, it still reflected meaningful improvements in balance and functional performance [9]. Furthermore, the interpretation should be contextualised: most BBS items are performed in an upright position, and the patient required both a KAFO and elbow crutches to maintain orthostasis due to the absence of active movement in the left LL. This limitation inherently restricted the scoring potential on the scale.

Despite the positive outcomes, certain limitations must be acknowledged. The effective duration of RAGT per session was shorter than current recommendations, which may have limited neuroplastic adaptation. Furthermore, although no adverse events or excessive fatigue were observed, the high cost and limited availability of exoskeleton devices remain significant barriers to widespread clinical implementation [4].

The integration of RAGT with an exoskeleton into a multidisciplinary paediatric rehabilitation programme proved to be a feasible and promising adjunct in the functional recovery of an adolescent with incomplete SCI due to idiopathic TM. In addition to improving spatiotemporal gait parameters, the intervention promoted more symmetrical weight distribution, reduced compensatory strategies, and enhanced gait balance, speed, and effort tolerance—key components for safe and efficient ambulation. These outcomes support the role of robotic technologies as valuable complements to conventional rehabilitation, particularly in paediatric populations where neuroplastic potential is high and long-term functional outcomes are critical. However, these findings are limited by the single-subject design. Kinematic analysis is also recommended to enhance the reliability of gait assessment beyond clinical observation. Further studies with larger cohorts and extended follow-up are warranted to confirm these preliminary results and to better understand the long-term effects of RAGT on neuroplasticity and autonomy.

5. Conclusions

This case supports the feasibility and therapeutic potential of integrating RAGT into paediatric neurorehabilitation. Clinically meaningful gains were observed in gait symmetry, weight transfer, and functional mobility. Despite limited training time and the single-case design, findings suggest exoskeletons may support broader motor goals beyond gait parameters. Further research is needed to explore long-term effects in promoting neuroplasticity and autonomy.