“Effects of robot-assisted gait training combined with traditional treatment in hereditary spastic paraplegia: An observational pilot cohort study” Authors: Eleonora Diella, Giorgia Malerba, Emilia Biffi, Roberta Morganti, Rossella Cima, Morena Delle Fave, Maria Grazia D’Angelo Published: 2026 ABSTRACT AbstractEvidence-based guidelines for motor rehabilitation in Hereditary Spastic Paraplegia (HSP) are still lacking but robotic gait training may offer benefits for motor function and walking. The aim of this study is to evaluate the effects of a robot-assisted gait training (RAGT) combined with traditional treatment (TT) on gait pattern, motor skills and quality of life in subjects affected by HSP. 40 patients (33 pure; 20 pediatric) underwent an intensive program with RAGT and TT daily (five-times a week for three consecutive weeks). Patients were evaluated pre and post intervention with gait analysis, gross-motor and timed walking tests and quality of life questionnaires. A three-month follow-up was conducted. A statistically significant improvement was observed in the Six-Minute Walk Test (6MWT). Enhancements have been obtained also in the Ten-Meter Walk Test (10mWT), the Berg Balance Scale (BBS) and the GMFM-88. The Spastic Paraplegia Rating Scale (SPRS) suggested reduction in disease severity. Quality of life’s questionnaires showed improved perceived physical health. Gait analysis data did not show significant changes. Intensive treatment with a combination of RAGT and TT improves functional skills and quality of life in HSP. These new findings could guide clinical management and development of future rehabilitation studies.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-56673-5.Keywords: Hereditary spastic paraplegia, Neuro-rehabilitation, Robot-assisted gait training, Intensive rehabilitationSubject terms: Diseases, Health care, Medical research, Neurology, Neuroscience Introduction Hereditary spastic paraplegia (HSP) is a rare disease (prevalence 0.1–9.6/100,000 around the world) that refers to a very heterogeneous group of neurodegenerative disorders in which the prominent clinical feature is gait abnormality mainly caused by spasticity of the lower limbs due to length-dependent damage to the first motor neuron1,2. HSP onset can occur at any age (pediatric and adult forms) and is characterized by progressive pyramidal symptoms mainly due to primary retrograde degeneration of the corticospinal axonal fibers3,4. From a clinical perspective, HSPs have been classified into “uncomplicated” and “complicated” forms, on the basis of additional features, besides pyramidal signs. “Pure” forms are characterized by lower limb spasticity and weakness, hyperreflexia, impaired proprioception and vibratory sensitivity. “Complicated” forms could present, in addition, cerebellar ataxia, peripheral neuropathy, retinopathy, epilepsy and cognitive impairment5. At present, no effective therapy can prevent, delay or reverse nerve degeneration in HSP and treatments are mainly focused on symptoms management. Physical therapy seems to be a core element to reduce functional gait impairment; however, conclusive evidence-based guidelines are not available yet6–9. Up to now, the effectiveness of physical therapy in patients with HSP is documented only by a small number of case reports and uncontrolled studies with quite limited evidence10–12. Furthermore, a crucial issue is the selection of outcome measures. Cubillos-Arcila has recently shown that the Spastic Paraplegia Rating Scale (SPRS) and Timed-Function Tests (i.e. Six-Minute Walk Test), even if valid instruments to assess disease severity, cannot capture the slow HSP motor evolution, showing the need of additional biomarkers (i.e. biomechanics of gait or neuroimaging)13. Furthermore, the absence of natural history data represents a major limit to study the impact of new physical or pharmacological treatments on the evolution of the disease. Recently, only a study presents prospective disease progression of HSP over a period of 4.5 years of follow-up14. Abnormal spastic gait is the cardinal feature of HSP. Three-dimensional clinical gait analysis objectively quantifies gait characteristics and is widely used to better understand compensatory mechanisms in pathological gait patterns. Concerning spatio-temporal parameters, HSP gait seems defined by reduced step length and increased step width and it is usually combined with reduced walking speed15,16. Common HSP gait kinematic pattern seems characterized by knee hyperextension, a reduced range of motion at the hip, knee and ankle in combination with a reduced foot lift12. Further, a large trunk tilt velocity during the loading response and the stance/swing transition is also reported17,18. In HSP, gait studies support a relevant heterogeneity of gait patterns, especially considering patients in different stages of the disease19–22. One of the most important areas of intervention in neuro-rehabilitation is the treatment of gait abnormalities. Robot-assisted gait training (RAGT) systems have been introduced in clinical practice to facilitate gait training in different neurological diseases. One of the most widely installed walking robots to deliver RAGT is the Lokomat (Hocoma AG, Volketswil, Switzerland) that is used for the rehabilitation of people with spinal cord injury, stroke, Parkinson’s disease, multiple sclerosis and cerebral palsy23–27. Theoretically, this kind of intensive RAGT training approach in HSP patients could lead to a combination of pathophysiological, biomechanical and reconditioning effects. Activity-dependent neuroplasticity of the spinal and supra-spinal control mechanisms of walking could enable the recovery of gait pattern, function and coordination by replicating a “normal” gait pattern using RAGT, further enhanced by the proprioceptive sensory inputs provided by the training28,29. Moreover, intensive RAGT training may be beneficial for gait biomechanics and function, allowing changes in speed, weight bearing control and balance [7; 30]. In addition, reduced physical activity, frequently reported and associated with impaired walking capacity in HSP patients30 may be overcome by RAGT training through the augmented physical activity induced by the repetitive intensive active robotic training31,32. To our knowledge, only two studies explored the potential benefits of a mechanical assisted locomotion training in patients with HSP. A case report presents the outcome of RAGT combined with physiotherapy in a 28-year-old man with a pure HSP, showing improvements in walking speed and balance, but not in gait kinematics and kinetics33. Furthermore, Bertolucci in 2015 conducted a six-week Lokomat training protocol on 13 patients with genetic defined HSP demonstrating improvements in balance and walking ability with a positive impact on quality of life7. This study aims to evaluate the effects of an intensive rehabilitation program combining Lokomat training with traditional therapy (TT) on gait pattern, motor function, and quality of life in ambulatory patients with HSP in a larger cohort, including both pediatric and adult individuals, with respect to previous studies. Furthermore, this study’s results enrich those obtained in previous papers by including a tridimensional gait analysis evaluation as secondary outcome measure and a 3-months follow-up assessment. We hypothesized that an intensive rehabilitative treatment could improve gait endurance and motor function of patients with HSP, with a positive impact on quality of life. Methods Design This was a retrospective observational pilot cohort study conducted following STROBE guidelines (Supplementary Material 1). The study scheme is represented in Fig. 1. Study flowchart. Participants Patients were recruited from those treated at the Scientific Institute Eugenio Medea in Bosisio Parini, Italy. We included both inpatients and outpatients seen between 2012 and 2024 who presented clinical signs of spastic paraplegia. They were all responding to the following inclusion criteria: genetic or clinical diagnosis of HSP, age above 4 years, femoral length > 23 cm, ability to ambulate indoors, with or without assistive devices. Exclusion criteria were: botulinum toxin injections or lower limb surgery within the previous 6 months, severe lower limb contractures, severe osteoporosis, unhealed skin lesions in the lower extremities, cardiovascular instability, acute or progressive neurologic disorders and behavioral problems. This study was performed in accordance with the Declaration of Helsinki and the Italian Pediatric National Ethics Committee approved the observational study protocol (protocol code: 0143318; date of approval: 7 November 2024). The trial was registered on clinicaltrials.gov (NCT06728787). Patients or their parents provided written informed consent. Study protocol The intensive rehabilitation protocol consisted of 15 sessions with the Lokomat (Hocoma AG, Volketswil, Switzerland), an active lower limb exoskeleton. Lokomat gait training was combined with 15 sessions of traditional treatment (TT). Sessions were performed twice daily (1 Lokomat + 1 TT), five-times a week (Monday-Friday) for three consecutive weeks. Both robotic training and TT were delivered by trained physiotherapists with extensive experience in the field of HSP. Assessments and treatments were carried out by different operators, following strictly a shared defined protocol. Lokomat sessions of 45 min started with the robot’s maximum assistance (guidance force 100%), body weight support (BWS) set at 50% and at a speed comfortable for the patient. Both guidance force (GF) and BWS were gradually reduced based on functional capacity, individual response to the intervention and the goal of optimizing training intensity, while always avoiding excessive fatigue. Throughout the treatment, exercises were modified in speed, duration and difficulty level to maintain high patient motivation during each session. A therapist was always present during the sessions to monitor the performance as well as to give feedback to the patient. For the traditional therapy (TT), a predefined exercise protocol was designed to address disease-specific targets. Each 45-minute session followed a multidimensional approach aimed at improving gait capacity. Sessions began with stretching exercises to reduce muscle spasticity and enhance mobility, followed by both selective and functional strength training to improve muscle control. The final part of the session focused on gait pattern training and both static and dynamic balance exercises, in line with motor learning principles that emphasize repetitive, task-specific practice with appropriate feedback18. Outcome measures Body Structure and Function, Activity and Participation, as defined by the International Classification of Disability and Health (ICF) were thoroughly explored using the outcome measures selected to detect limitations at different levels of functioning34. The primary outcome was the Six-Minute Walk Test (6MWT) that assesses endurance during self-paced walk by measuring the distance walked within 6 min along a standardized 25-meters path. Patients were instructed to walk as far as possible and were constantly encouraged by the therapist35. Secondary outcomes included: the Ten-Meters Walk Test (10mWT) that records walking speed for a 10-meter distance36, the Ten-Meters Walk Test (10mWT) that records walking speed for a 10-meter distance36, the Spastic Paraplegia Rating Scale (SPRS) as an indicator of disease severity and progression37, the Spastic Paraplegia Rating Scale (SPRS) as an indicator of disease severity and progression37, the Gillette Functional Assessment Questionnaire (FAQ) as an indicator of the level of walking ability38, the Gillette Functional Assessment Questionnaire (FAQ) as an indicator of the level of walking ability38, the Berg Balance Scale (BBS), a 14-item tool that evaluates static and dynamic balance on a 5-point ordinal scale39(only for adult patients), the Berg Balance Scale (BBS), a 14-item tool that evaluates static and dynamic balance on a 5-point ordinal scale39(only for adult patients), the Gross Motor Function Measure-88 (GMFM-88), that consists of 88 items to assess functional abilities divided into the following sections: (A) lying and rolling, (B) sitting, (C) crawling and kneeling, (D) standing, and (E) walking, running, and jumping40(only for pediatric patients). the Gross Motor Function Measure-88 (GMFM-88), that consists of 88 items to assess functional abilities divided into the following sections: (A) lying and rolling, (B) sitting, (C) crawling and kneeling, (D) standing, and (E) walking, running, and jumping40(only for pediatric patients). A quantitative analysis of the gait was performed in the 3D gait-analysis laboratory. The laboratory is equipped with eight optoelectronic cameras, an optoelectronic system (Elite, BTS Bioengineering, Milan, Italy) with a sampling rate of 100 Hz, and two force plates (Kistler Group, Winterthur, Switzerland) embedded in the floor. All measurements and analysis were performed according to the Davis protocol41. After the application of the markers directly on the patient’s skin, each subject was asked to walk barefoot along a 10-meter walkway, for at least five trials. Personal perspective was assessed using two patient-reported outcome measures (PROMs) which captured the patient’s point of view about symptoms severity, functional and psychological challenges and quality of life. The Medical Outcome Survey Short Form (SF-36) includes 36 items aggregated into two summary measures, i.e. physical component and mental component of health, according to Apolone et al.42. The Hereditary Spastic Paraplegia-Self Notion And Perception questionnaire (HSP-SNAP) is a disease-specific tool designed to assess personal perception on specific symptoms related to HSP such as stiffness, weakness, imbalance, reduced endurance, fatigue and pain43. Patients were evaluated before the beginning of the treatment (T0), at the end (T1) and after a 3-month follow-up period (T2). Gait parameters extraction For each patient, an expert physiotherapist collected and processed at least five trials for the left and the right limbs using dedicated software (EliteClinic, BTS Bioengineering, Milan, Italy). In line with the routine clinical workflow under which these retrospective data were collected, one representative trial was selected for further analysis by an expert physiotherapist, based on overall technical quality and consistency with the patient’s habitual gait pattern. For the selected trial, one gait cycle per limb was considered. BTS Smart Clinic software was used to extract spatio-temporal data, and kinematic and kinetic curves for each selected gait cycle. Basic spatio-temporal gait parameters directly extracted from BTS Smart Clinic included: Walking velocity [m/s]. Walking velocity [m/s]. Stance phase [% of gait cycle]. Stance phase [% of gait cycle]. Length and width of the step [m]. Length and width of the step [m]. An ad hoc algorithm was developed in Matlab (The Mathworks, Natick, MA, USA) to process the exported kinematic and kinetic curves and compute the following summary parameters. Kinetics: Maximum values of the ankle, knee and hip moment in the sagittal plane [Nm/kg]. Maximum values of the ankle, knee and hip moment in the sagittal plane [Nm/kg]. Minimum values of the knee and hip moment in the sagittal plane [Nm/kg]. Minimum values of the knee and hip moment in the sagittal plane [Nm/kg]. Maximum values of the ankle, knee and hip power in the sagittal plane [W/kg]. Maximum values of the ankle, knee and hip power in the sagittal plane [W/kg]. Minimum values of the ankle power in the sagittal plane [W/kg]. Minimum values of the ankle power in the sagittal plane [W/kg]. Kinematics: Mean value of the foot progression in stance [degrees]. Mean value of the foot progression in stance [degrees]. Maximum and minimum flexion angle, the range of motion (ROM) and the angle at initial contact in the sagittal plane for ankle, knee, and hip [degrees]. Maximum and minimum flexion angle, the range of motion (ROM) and the angle at initial contact in the sagittal plane for ankle, knee, and hip [degrees]. ROM of pelvic tilt, obliquity and rotation [degrees]. ROM of pelvic tilt, obliquity and rotation [degrees]. All parameters were then merged into a single database for further analysis. The mean value between left and right side was considered for all the parameters. Statistics Not all patients were able to complete the T2 assessments due to personal and logistical reasons, as T2 was scheduled three months after the end of the intensive rehabilitation (T1). Therefore, the analysis was conducted on two separate datasets. First, to exploit the whole sample, we performed a T0 vs. T1 data analysis. Then, to exploit the presence of follow-up assessments, we performed a repeated measure (T0, T1, T2) analysis. Data distribution was assessed using the Shapiro-Wilk test. If data had normal distribution, they were presented as mean ±standard deviation (SD) and parametric tests were chosen. When normality was not verified, non-parametric tests were used, and data were represented with median and interquartile range (IQR) values. For the first analysis, the Paired t-test or the non-parametric Wilcoxon signed-rank test were applied to analyze the statistical difference between baseline (T0) and post-treatment (T1). Significance was set at p < 0.05. For the baseline versus post-treatment analysis, change scores were defined as the difference between post-treatment and baseline values (Δ = T1 − T0). For the second analysis, repeated measures ANOVA test or the non-parametric Friedman test were applied to analyze the statistical difference between baseline (T0), post-treatment (T1) and follow-up (T2) scores. The paired t-test or the non-parametric Wilcoxon test with Bonferroni correction were applied as post hoc analysis on the significant results of the ANOVA and Friedman test, considering a p-value corrected (p = 0.017). The statistical analysis was performed with SPSS v21. Results Participants We recruited 40 patients (mean age = 23.85 years; SD = 18.41 years; age range [5.31–66.02], 26 males; 20 pediatric) affected by pure (33/40) and complicated (7/40) forms of HSP. Twenty-five out of forty patients had a confirmed genetic diagnosis of HSP. In the remaining fifteen, who presented clinical signs of spastic paraplegia, other causes of spasticity were excluded through brain and spinal cord imaging, metabolic analysis, clinical history (i.e. pregnancy and/or delivery complication, Apgar evaluation at birth) (Table 1). Table 1Patients’ characteristics of the sample recruited. P=pure form. C=complicated form. SPG: spastic paraplegia gene. CD: clinical diagnosis M: male; F: female. Follow-up: Y: yes; N: no.PatientGenderAge (Years)Disease Onset(Years Old)Genetics / Clinical DiagnosisP/CFollow-upP1M7.241SPG54CNP2M5.311CDCYP3M9.691SPG3APNP4M18.991CDPNP5M5.351SPG3APNP6M20.051CDPNP7M13.302SPG3APNP8M8.761CDCNP9M18.9814SPG11CYP10M41.5921CDPYP11F6.721CDPYP12F21.566CDPYP13M10.281SPG3APNP14F27.483SPG30PYP15F9.082SPG3APYP16M21.471CDPNP17M13.022CDPNP18F16.683CDPNP19F10.792SPG3APNP20F11.782SPG3APNP21M6.351SPG4PYP22F63.5142CDPYP23F5.642SPG4PYP24M23.284SPG31PYP25M52.3342SPG4PNP26M25.574SPG31PYP27F64.6442CDPNP28M7.641SPG3APNP29M14.961CDPNP30F6.892SPG4PNP31F42.6637SPG4PYP32M52.9036SPG7CYP33M24.7114SPG11CNP34M55.7730SPG7CYP35F32.723SPG30PYP36M24.994SPG31PNP37F66.024SPG31PYP38M16.221CDPYP39M14.571CDPNP40M54.3442SPG4PN Patients’ characteristics of the sample recruited. P=pure form. C=complicated form. SPG: spastic paraplegia gene. CD: clinical diagnosis M: male; F: female. Follow-up: Y: yes; N: no. Baseline vs. post-treatment analysis Clinical assessments Test results were collected from 40 patients, with variations in the number of participants for each metric, mainly due to differences in age at evaluation, as some outcomes are specific to certain age ranges. In particular, the GMFM-88 was administered exclusively to patients under 18 years of age; the BBS only on adult subjects; the SPRS and the HSP-SNAP were collected on patients over 9 years old and, the SF-36 was filled out only by patients over 14 years old44. Figure 2 presents the results of the clinical assessments that showed significant differences across the T0 and T1, as determined by the Paired t-test and the non-parametric Wilcoxon signed-rank test. All the data are reported in Supplementary Material 2 (Table 1). Results of the clinical assessments T0 vs. T1. Patients’ performance in the 6MWT significantly increased from a mean of 339.71 m at T0 to 370.88 m at T1, indicating improved walking endurance. The 10mWT also showed a significant improvement, with the median time decreasing from 7.89 s (1.27 m/s) at T0 to 6.74 s (1.48 m/s) at T1, reflecting increased walking speed. No significant differences for the FAQ values were observed. The BBS score significantly improved from a median of 47 at T0 to 49 at T1, demonstrating enhanced postural stability. The SPRS score significantly decreased from a mean of 16.7 at T0 to 15.5 at T1, suggesting a reduction in the severity of HSP symptoms. The SF-36 (Physical Component Summary) score significantly increased from a mean of 39.1 at T0 to 41 at T1, indicating better perceived physical health. The HSP-SNAP score also showed a significant increase, rising from a median of 20.8 at T0 to 26.8 at T1, suggesting improvement in personal perception of symptoms related to walking. The median value for GMFM-88 remained at 88% but the change between T0 and T1 was statistically significant. Similarly, GMFM-C and GMFM-D showed significant changes. GMFM-E significantly increased from a mean of 64% at T0 to 66% at T1, showing progress in advanced motor abilities. Gait parameters Gait analysis assessment was performed by 35 patients. No gait parameter showed statistically significant differences between T0 and T1. All the data are reported in Supplementary Material 2 (Table 2). Baseline vs. post-treatment vs. follow-up analysis Clinical assessments To assess the long-term effects of the treatment, a follow-up analysis was conducted on the 18 subjects who completed all phases of the study (follow-up at 3 months). Supplementary Material 2 (Table 3) presents the results of the repeated measures ANOVA or Friedman test, with post hoc analysis, used to compare baseline (T0), post-treatment (T1), and follow-up (T2) data. Figure 3 presents the results of the clinical assessments that showed significant differences across the three time points (T0, T1, and T2). The 6MWT showed a significant improvement from T0 to T1, indicating increased walking endurance, and this improvement was maintained at T2, demonstrating a lasting effect of the treatment on walking resistance. The BBS revealed a significant improvement from T0 to T1 and stability of the value at T2, indicating a long-term enhancement in balance capabilities. The SPRS indicated a significant reduction between T0 and T1; however, at follow-up, a significant increase was observed from T1 to T2, returning to values close to pre-treatment levels. The HSP-SNAP score showed a significant increase from T0 to T1; although a slight reduction was observed at T2, scores remained higher than the baseline. Both the GMFM-88 and GMFM-E scales showed significant overall improvements in the ANOVA, indicating enhanced gross motor function. However, post hoc analysis did not reveal significant differences between individual time points. Results of the clinical assessments T0 vs. T1 vs. T2. Gait parameters T2 gait analysis assessment was performed on 15 patients. Supplementary Material 2 (Table 4) shows the results of the repeated measures ANOVA test with post hoc analysis, comparing T0, T1, and T2 spatio-temporal gait parameters. However, no statistically significant differences were found across the three time points. Discussion Motor rehabilitation is a key component in the lifelong management of patients with hereditary spastic paraplegia. Gait training should be one of the primary focus of the treatment in patients who retain walking ability and autonomy. Building on previous literature highlighting the benefits of robotic gait training and leveraging a larger cohort of patients and additional assessments (gait pattern) and time-points (follow up assessment), this study evaluated the effects of intensive robot-assisted gait training combined with traditional treatment on gait pattern, motor function and quality of life in individuals with HSP. Concerning functional motor skills, a significant improvement emerged in the primary outcome (∆ 6MWT = 31.17 m) at the end of the training which is bigger than the minimum clinically important difference (MCID) reported by Storm and collaborators45, underlining the positive clinical impact of this result. Enrolled patients demonstrated enhanced gait endurance, although the mean improvement in the 6MWT was smaller (approximately 30 m) compared to the findings of Bertolucci et al. (mean variation of 40 m). This difference is likely due to the subjects recruited in this study that had higher 6MWT scores at T0 (indicating milder symptoms), which left a smaller range for improvement. At follow-up, 6MWT values showed general stability, demonstrating good retention of endurance improvement over time. The gain in gait resistance and its maintenance are meaningful results, as this is considered a crucial issue by both patients and clinicians46,47. Data collected in this work showed also a significant improvement in gait speed, that increase of 0.21 m/s as assessed by the 10-Meter Walk Test (10mWT) at T1 in line with what was previously found by Bertolucci and Seo7,33. Considering a MCID of 0.1 m/s for the 10mWT48, these results can be considered clinically meaningful. Conversely, the FAQ did not show any changes, likely because it may not be sensitive enough to detect minimal changes in this chronic, slowly evolving disease. We observed a statistically significant pre–post change in the Berg Balance Scale (BBS), with an average improvement of about 2 points. Although this value is slightly below the MCID of 3 points reported by Gervasoni et al.48, it still represents a positive change in balance control. These constructs directly correlate with fear and risk of falling that are frequent impairments in HSP patients and the retention of BBS improvement at the follow-up highlights the importance of this result49. The severity of the disease, recorded with the SPRS, showed a significant pre-post change with a reduction of 1.2 point in the total score. However, the follow-up values go back to T0, showing no retention of the improvements obtained at the end of the treatment. GMFM-88 used in the pediatric subsample showed statistically significant improvements primarily in dimensions D (standing) and E (walking, running, and jumping), with mean changes of 1% and 2% respectively. Considering the reported MCID ranges for GMFM dimensions D and E45, 0.8–5.2% and 0.3–4.9% respectively, these changes fall within the expected range and can therefore be considered clinically meaningful. Gross-motor improvements in children were expected50: assisted gait training in addition to traditional treatment, working on walking pattern and its ecological transfer in real setting, can lead to improvements in static and dynamic gait control with positive impacts in correlated areas of function. The small number of GMFM-88 evaluations in the follow-up (5 subjects) prevent from drawing conclusive considerations on the retention of the training effect on gross-motor skills. At the end of the treatment, SF-36 physical dimension and HSP-SNAP scores showed statistically significant change. Specifically, the HSP-SNAP, a disease-specific questionnaire that assesses patients’ perception of key symptoms such as stiffness, weakness, reduced endurance, unbalance, fatigue and pain, recorded a reduction of 6 points. In contrast, the SF-36, a widely validated and used general PROM that is not specific to HSP patients, showed lower changes. This may be attributed to its lower specificity and sensitivity to changes compared to the HSP-SNAP, which is specifically tailored to this disease51. SF-36 results differ from those reported by Bertolucci and colleagues7, who observed a substantial improvement in both physical and mental dimensions in the pre-post evaluation, however, they used a different score computation method from the one proposed by Apolone42. At the follow-up, HSP-SNAP values decreased but did not return to baseline, indicating partial maintenance of the perceived personal well-being. No changes in gait analysis data were observed. These results must be discussed considering the high variability of the gait pattern of HSP patients (adult and pediatric) as underlined by the literature12,19,21. Wolf et al. studied a group of 29 subjects with HSP, whose age was comparable to that of our patients, and clustered patients in five subgroups according to their gait features: crouch gait, recurvatum gait, stiff knee, jump knee and “norm-like” pattern17. This gait variability might have blurred major changes flattening data and hiding statistically significant improvements. Further studies with a more homogeneous population in terms of similar gait pattern might clarify this critical issue and potentially support the use of Lokomat training for walking pattern improvement in patients with HSP. All subjects showed good compliance during the training, no collateral effects were reported and spasticity was well managed. Some patients reported increased fatigability during the final week of training, which required adjustments of the intensity of the individual sessions. This intensive RAGT approach in HSP patients seems to induce biomechanical and reconditioning effects that may be interconnected with pathophysiological ones. However, the data collected so far are not able to disentangle which mechanism (i.e. compensation to the length dependent distal degeneration of the axon of the first motor neuron and/or neuroplasticity change) underlies the observed changes. This work has some limitations. First of all, patients have genetic and clinical heterogeneity and a wide range of age (5–63 years) that was due to the observational nature of the study. Nevertheless, this study design allowed us to reach a significant number of participants and to explore the effects of RAGT in subjects with different characteristics. Another limitation concerns gait analysis processing. In line with the routine clinical practice under which these retrospective data were collected, one representative trial, including one gait cycle per limb, was selected for each patient rather than averaging multiple accepted gait cycles. Although this approach reflects standard clinical reporting, it may not fully capture intra-subject gait variability. Future studies should consider averaging multiple gait cycles to provide a more comprehensive description of gait pattern. The study design and its retrospective nature are other limitations of the study, in addition to the reduced number of patients that completed the follow-up assessment. Given the lack of a control group, our findings cannot be generalized and should be interpreted with caution. The efficacy of the proposed program needs to be confirmed through randomized controlled trials with larger samples and long-term follow-up. Nevertheless, despite these limitations, these results should be taken into account for further studies, considering the HSP’s rarity, that no larger group of HSP patients have been studied before and that this is the first rehabilitative study that also includes the pediatric population. Future research could explore the use of alternative outcome measures as primary endpoints and/or adopt different gait analysis methods. For example, analyzing all steps during walking - rather than focusing solely on the central step, as done by Pulido-Valdeolivas et al.52 may provide a more comprehensive and realistic representation of patients’ gait patterns. Among potential outcome measures, the Timed Up and Go (TUG) test appears to be particularly sensitive in the HSP patient population. Cubillos Arcila et al.14 recommend selecting the TUG at maximal velocity as the primary endpoint in future clinical trials, as it may require the smallest sample size to detect meaningful clinical changes. Conclusion This work shows that a 3-week intensive treatment with a combination of RAGT and traditional treatment is well tolerated and can improve functional skills and quality of life in patients affected by HSP. Changes in gait pattern need to be further investigated considering more homogeneous groups of subjects. These new findings could guide clinical management and development of future rehabilitation studies in this rare disease. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Acknowledgements We are grateful to all the patients and their family for the constant collaboration. We also thank ViPS ETS association for the support. Author contributions ED, MGD, RC: Conceptualization - ED, GM, EB: Data Curation - GM, EB: Formal Analysis - ED, RM, MDF: Investigation - MGD, RC: Supervision - ED, GM: Writing –original draft - ED, GM, EB, MGD: Writing, review and editing. All authors read and approved the final manuscript. Funding This study was supported by the Italian Ministry of Health (Ricerca Corrente 2024/2026 to E. B.) and by the Italian Ministry of University and Research under the complementary actions to the NRRP “Fit4MedRob - Fit for Medical Robotics” Grant (# PNC0000007). Data availability The data that support the findings of this study are openly available on Zenodo at: https://doi.org/10.5281/zenodo.15674322. Declarations Competing interests The authors declare no competing interests. Ethics approval This study was performed in accordance with the Declaration of Helsinki and the Italian Pediatric National Ethics Committee approved the observational study protocol (protocol code: 0143318; date of approval: 7 November 2024). The trial was registered on clinicaltrials.gov (NCT06728787). Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement The data that support the findings of this study are openly available on Zenodo at: https://doi.org/10.5281/zenodo.15674322.