Background and Purpose: This case report describes the effects of locomotor training using body-weight support (BWS) on a treadmill and during overground walking on mobility in a child with severe cerebellar ataxia who was nonambulatory. To date, no studies have examined the efficacy of this intervention in people with cerebellar ataxia.

Case Description: The patient was a 13-year-old girl who had a cerebellar/brainstem infarct 16 months before the intervention. Her long-term goal was to walk independently in her home with a walker.

Intervention: Locomotor training using a BWS system both on the treadmill and during overground walking was implemented 5 days a week for 4 weeks in a clinic. Locomotor training using BWS on a treadmill was continued 5 days a week for 4 months at home.

Outcomes: Prior to training, she was able to take steps on her own with the help of another person, but did not take full weight on her feet or walk on a regular basis. At 6 months, she walked for household distances. Prior to training, her Pediatric Functional Independence Measure scores were 3 (moderate assistance) for all transfers, 2 (maximal assistance) for walking, and 1 (total assistance) for stairs. At 6 months, her scores were 6 (modified independence) for transfers, 5 (supervision) for walking, and 4 (minimal assistance) for stairs. Prior to training, she was unable to take independent steps during treadmill walking; at 6 months, all of her steps were unassisted.

Discussion and Conclusion: Locomotor training using BWS on a treadmill in conjunction with overground gait training may be an effective way to improve ambulatory function in individuals with severe cerebellar ataxia, but the intensity and duration of training required for functionally significant improvements may be prolonged.

One of the most commonly stated goals in the rehabilitation of populations with neurologic problems is the recovery of walking capacity.14 Although there is no consensus currently on the best way to train safe, efficient, and independent walking, evidence exists that task-specific gait training can have more beneficial effects on functional gait outcomes than nonspecific rehabilitation approaches.5,6 Locomotor training using body-weight support on a treadmill (BWST) is an example of task-specific gait training that uses a harness to provide partial body-weight support in conjunction with a motorized treadmill. Researchers examining locomotor training have suggested that training using BWST has several advantages: (1) the body-weight support harness allows a progressive increase in the demands for postural control, (2) the treadmill allows systematic control and progression of the speed at which walking is performed, and (3) the repetitive training of a complete gait cycle enables a more appropriate pattern of sensory input associated with the different phases of gait to stimulate the locomotor pattern.79 In addition, locomotor training using BWST allows the therapist to provide manual assistance to help the patient simulate a more normal walking pattern.

Previous research has presented mixed results regarding the efficacy of locomotor training using BWST in patients with a variety of diagnoses, including stroke,1,5,10 spinal cord injury,11,12 and cerebral palsy.13 A Cochrane review by Moseley et al10 investigated the results of gait training using BWST after stroke in 11 randomized controlled trials involving 458 participants and reported no statistically significant differences between gait training using a treadmill with or without body-weight support and other interventions for walking speed and dependency. Matching the type of locomotor training to specific patient characteristics, such as gait speed, may be required to optimize outcomes. For example, patients who have had a stroke and walk slower than 0.4 m/s benefited most from body-weight support and treadmill training, whereas those who walked faster than 0.4 m/s required the addition of overground training to training using BWST.1

Guidelines regarding optimal ways to combine overground training and training using BWST in specific patient subgroups are still being developed. Helbostad14 suggested that, despite the mixed evidence for the efficacy of training using BWST for improving gait in patients who are ambulatory, it may be the only alternative for ambulation training in people who are unable or require significant assistance to walk.

Disordered or ataxic gait is a defining characteristic of cerebellar pathology. Ataxic gait has been characterized by a widened or alternatively variable base of support, inappropriate timing of foot placement and reduced step frequency, increased step width, and prolonged time in double-limb support.15 Both impaired postural stability and decomposition of multijoint leg movements appear to be factors in cerebellar gait ataxia.16,17 This combination of impaired balance and dyscoordination in lower-limb dynamics suggests a strong rationale for the use of locomotor training using BWST in people with ataxic gait. However, BWST uses unweighting during locomotor training, and this unweighting is in contrast to previous authors18 who have recommended the use of weights on the trunk and limbs to decrease ataxic movements in patients with cerebellar pathology.

In addition to their role in the control of balance and gait, cerebellar structures also are important in practice-dependent motor adaptation and learning in many different systems, including those controlling locomotion.16,17 Damage to the cerebellum affects the extent and rate at which individuals adapt locomotion to new contexts.16 This suggests that the training of individuals with cerebellar pathology may require a longer duration or intensity of practice to improve locomotor function regardless of the type of training.

The purpose of this case report is to describe the effect of locomotor training using BWST in conjunction with overground walking training on mobility function in a 13-year-old patient with severe cerebellar ataxia.

Case Description

Patient Description

The patient was a 13-year-old girl who had been healthy previously and who had a spontaneous posterior fossa hemorrhage with resultant cerebellar and brainstem infarcts. She was initially unresponsive and on a ventilator. She regained consciousness after approximately 11 weeks. In addition to rendering her nonambulatory, initial sequelae from the hemorrhage included ataxia, decreased coordination, weakness, and multiple cranial nerve palsies.

This individual was selected for the case report involving locomotor training using BWST for several reasons: one of her goals for therapy was to improve her ability to walk independently, she was limited in her ability to participate in mobility training because she required maximum assistance of 2 people to walk, she had a brief exposure (less than 10% of total treatment time) to locomotor training using BWST during her inpatient rehabilitation stay and was receptive to it, and her family was very supportive and willing to transport her to the clinic for physical therapy every day. In addition, during her inpatient stay, weights were tried on both limbs and the trunk to reduce ataxic movements, with no appreciable gains in either functional limb movements or improved postural stability in sitting, standing, or walking. The patient began the intervention outlined in this case report 16 months after her injury.


A physical therapist performed the tests and measures before the beginning of clinic training, immediately after completion of clinic training, 1 month after completion of clinic training, and after the completion of 4 months of home training. At the beginning of the intervention, (16 months after the hemorrhage), our patient required a tracheotomy and feeding tube, she was dependent in all activities of daily living, and she used a wheelchair for in-home and community mobility.

Table 1 summarizes her physical impairments at the initiation of training. She exhibited significant ataxia during both upper- and lower-extremity movements, including dysmetria when reaching for objects, dysdiadochokinesia when performing rapid alternating movements, weakness (4 out of 5 on the manual muscle test),19 spasticity greater on the right side (score of 2 on the Modified Ashworth Scale),20 clonus, and a tight Achilles tendon on the right lower extremity. She also demonstrated multiple cranial nerve palsies (left vocal cord paresis, facial nerve palsies, and swallowing dysfunction) and sensory and perceptual deficits primarily on the left side, including decreased vision, decreased proprioception, and impaired hearing.

Table 1.

Impairments in Body Structure and Function as Measured at Baseline

At the beginning of the intervention, she required the assistance of 2 people to ambulate because of an inability to maintain upright posture and to produce an appropriate locomotor pattern in her legs. During gait, she exhibited poorly coordinated leg movements, which resulted in abnormal and variable swing foot trajectories and foot placement, increased variability in length and timing of steps, slow walking speed, profound trunk sway, and an inconsistent base of support (alternating too narrow or too wide). Because of the severity of her gait impairment, locomotor training using BWST was considered the only feasible method for this patient to safely and consistently practice walking. In addition, body-weight support allowed the therapists to provide manual assistance with stepping rather than having to assist with weight support and postural stability.

Her physical therapy diagnosis was “Impaired Motor Function and Sensory Integrity Associated With Nonprogressive Disorders of the Central Nervous System—Acquired in Adolescence or Adulthood” (Preferred Physical Therapist Practice Pattern 5D).21

Two measures of mobility were used as outcome measures: the Pediatric Functional Independence Measure (WeeFIM), specifically the transfers and locomotion subscales,22,23 and the Gillette Functional Walking Scale.24 The WeeFIM is a modified version of the Functional Independence Measure designed to indicate severity of disability based on the concept of “burden of care,” which refers to the amount of assistance required to perform a given activity. The WeeFIM has 18 items in 6 subscales; however, only the transfers and locomotion subscales were considered as outcome measures for this case report. We chose this test for its ability to monitor change over time in children with chronic disabilities. Ottenbacher et al23 found the responsiveness of the WeeFIM to be statistically significant (P<.05) for detecting change in functional abilities in 174 children with chronic disabilities over a 1-year period. Interrater reliability is high (intraclass correlation coefficient [ICC]=.90–.99), as is test-retest reliability (ICC=.98–.99).22 Ottenbacher et al22 also demonstrated that the ICC was the highest (.99) for the transfers and locomotion subscales that we used with our patient.

The Gillette Functional Walking Scale (the walking-scale portion of the Gillette Functional Assessment) is a 10-level scale that includes the entire range of walking abilities from nonambulatory to completely independent ambulation in all community settings and terrains. This scale was chosen for its ability to help document current functional status and for its ability to evaluate change as a result of interventions, specifically to track an individual's progress in establishing more independent gait. It is a reliable tool both between and within raters for a range of community ambulation over a time frame of 1 to 6 months.24 Good test-retest (ICC=.92), intrarater (ICC=.92), and interrater (ICC=.81) reliability have been demonstrated, and content and concurrent validity (assessed using Pearson correlations) also are significant at the .01 level.24

Mobility function also was measured during overground walking and during treadmill walking. Distance and level of assistance required were used to evaluate overground mobility function. In addition, the number of independent steps the patient was able to take while walking on the treadmill was measured weekly during the clinic training and then again at 1 and 5 months after completion of the clinic training.


The patient received locomotor training using BWST on a LiteGait system (model LGI 500)25,* (Fig. 1) in conjunction with practice in overground walking, 5 times a week for 4 weeks in the clinic. This was followed by 4 months of training using BWST in the home. At the conclusion of the 4-week intervention in the clinic, the patient had no locomotor training using BWST for 1 month because of a delay in locating a suitable in-home treadmill system. Factors that affected the selection of a home unit included limited space that required a treadmill that could be folded up and a treadmill that could be used at a speed of <0.2 m/s.

Figure 1.

Body-weight–supported treadmill training system used in the clinic.

The family purchased a Vision Fitness treadmill, and a Pneumex overhead frame and harness (model 07PW-1). The Pneumex harness needed to be altered because of the patient's discomfort with the leg attachments to the support harness. A groin piece was ordered from LiteGait and altered to fit the Pneumex harness to achieve the desired unweighting needed for training. Once her home treadmill system was installed (Fig. 2), she continued her locomotor training with BWST and overground training for 30 minutes a day, 5 days per week, with the assistance of a home care physical therapist or a trained rehabilitation aide.

Figure 2.

Body-weight–supported treadmill training system used in the home.

In the clinic, each 1-hour treatment session consisted of 15 minutes of locomotor training using BWST, followed by 15 to 20 minutes of overground walking training using the BWS harness. She was seen for a total of 19 sessions, and missed one treatment for surgical revision of her feeding tube. At the beginning of each session, the patient donned a harness with assistance and was brought into the standing position and assisted onto the treadmill.

Training Parameters and Progression

The parameters of training that were progressed with treatment are described below and summarized in Table 2.

Table 2.

Parameters for Training on a Treadmill and Overground Training With Body-Weight Supporta

Training time.

Initially, the patient attempted to ambulate on the treadmill for 20 minutes with 2 rest breaks. As she fatigued, her step control deteriorated and she required greater assistance; therefore, 15 minutes was set as the maximum time for treadmill walking for the remaining sessions. The gait training recommendations issued by the manufacturer of the body-weight–supported treadmill training system25 were followed, including keeping each treadmill training session to 15 minutes or less, ensuring that the patient took most of the steps with correct patterns, and ending each session when the therapist or patient fatigued or the patient's gait deteriorated. By the end of training, this patient was able to walk on the treadmill for 15 minutes with no rests. She was initially able to tolerate only 5 minutes of overground walking with body-weight support but progressed to 15 to 20 minutes with gradually decreasing percentage of support.

Body-weight support.

The level of body-weight support (determined by using a scale to measure body weight while standing on the treadmill in the harness) was adjusted across sessions to maximize upright posture and independent stepping. Initially, the harness provided approximately 30% of body-weight support for walking on the treadmill and during overground walking practice. This was progressively decreased to 10% of body-weight support by the second month of training.

Manual assistance.

Amount and type of manual assistance was decreased across training sessions. Initially, 3 physical therapists were required to provide manual assistance to the patient while walking on the treadmill in the support harness. One physical therapist assisted with each lower leg by placing one hand at the top of each foot to facilitate toe clearance at swing phase and heel-strike at initial stance and the other hand behind the knees to prevent knee hyperextension. As shown in Figure 1, the therapists providing manual assistance to control lower-limb placement either squatted or sat on the edge of the treadmill apparatus or on a low stool. A third assistant stabilized the patient at the hips and manually assisted with pelvic rotation. The patient was initially instructed to hold lightly onto handrails to provide additional support and upright posture on the treadmill as needed. Upper-extremity support was discontinued 2 months after the start of training, when the patient was able to maintain a vertical trunk position without the assistance of her arms; however, she still required manual assistance from the therapist at the hips. This assistance at the hips was maintained for almost 4 months because of her severe ataxia of the trunk.

Training speed.

Based on research with a stroke population that showed the efficacy of increased training speed on locomotor outcomes,26 our goal was to train gait at the fastest speed manageable by the patient. However, Hidler27 suggested that training speeds need to be set according to each person's ability to retain adequate motor control of his or her legs. At the beginning of training, our patient required 3 people to help her walk at a treadmill speed of 0.18 m/s with approximately 30% of body-weight support. We tried to increase her training speed, but found that if we attempted to increase treadmill speed faster than 0.31 m/s, our patient was not able to retain adequate motor control and stepping ability. We were not able to increase speed after the second week in the clinic, and the patient consistently trained at approximately 0.3 m/s.

Visual cues for foot placement.

Because of impaired lower-extremity proprioception, the patient used visual cues to guide foot placement during walking; however, this led to a forward head and trunk posture that decreased her postural stability. We alternated practicing stepping with visual guidance (eg, looking at her feet while walking) with looking straight ahead (increased use of proprioceptive and kinesthetic cues for foot placement). The amount of time spent in visually guided foot placement was gradually decreased, allowing the patient to gradually increase the number of steps she was able to take independently without visual guidance for foot placement.

Progression of training parameters involved first reducing the amount of manual assistance of the legs provided by therapists to improve stepping ability, then reducing body-weight support while maintaining independent stepping ability, and finally increasing speed of training while maintaining both postural stability and independent stepping ability. Concomitant to the progression of treadmill training parameters, we increased the time and duration of overground walking training, progressively decreased body-weight support in the harness, and increased gait speed during overground walking. The decision to prioritize independent stepping ability (decreased manual assistance) over postural stability (decreased body weight in the harness) was twofold: first and foremost, the patient herself was strongly committed to moving her limbs independently, and, second, manually advancing the limbs was physically taxing and fatiguing to the therapists. As the patient began to master independent stepping, body-weight support was reduced. Each session she attempted to ambulate by loading the maximum weight she was able to support and still maintain adequate stepping ability. Training time on the treadmill was consistently set at 15 minutes, but time spent on walking overground was increased as the patient developed stepping ability and tolerance for physical activity.

Vital signs were monitored daily in her home by a registered nurse and were stable, so they were not taken in the clinic. Oxygen saturation levels were taken when the patient showed any signs of respiratory distress, and they did not go below 90%.

Our patient also received in-home physical therapy twice a week for 90-minute sessions from the time of her discharge home from the rehabilitation center 6 months prior to the clinic intervention, during the 4-week session of training using BWST in the clinic, and for at least a year afterward. The goals of her in-home physical therapy program were to increase strength and coordination, to improve her ability to maintain balance in sitting and standing, and to increase independence in transfers. Specific interventions are summarized in the Appendix.


Table 3 compares scores on the 3 primary outcome measures over the course of the intervention. Prior to training, her score on the Gillette Functional Walking Scale was 2 (can do some stepping on her own with the help of another person, but does not take full weight on her feet or walk on a regular basis) and remained at 2 at the 4-week mark (ie, when the clinic training was completed) and at retention test 1 month after completion of the clinic training. Six months after initiation of the intervention, her score improved to 6 (walks for household distances).

Table 3.

Intervention Outcomes

Prior to training, her score on the WeeFIM was 3 (moderate assistance) for transfers and 2 (maximum assistance) for walking. At 4 weeks and at the retention test 1 month after the completion of the clinic training, her transfer score improved to 4 (minimal assistance), but her walking score did not change. Six months after the initiation of the intervention, her score was 6 (modified independence) for transfers and 5 (supervision) for walking.

The number of independent steps taken during treadmill walking was 0 prior to training, and improved weekly to 128 steps at 4 weeks and to 200 steps at the1-month follow-up after completion of the clinic training. Six months after initiation of the intervention, all steps on the treadmill were unassisted.

As noted in Table 2, there was a steady improvement in all training parameters for the training using BWST, including: an increase in gait speed from 0.18 to 0.31 m/s on the treadmill, increased endurance as shown by fewer rest breaks during the 15-minute treadmill sessions, decreased body-weight support from approximately 30% to 15%, and decreased level of assistance required for walking on the treadmill from 3 people to 1 person.

The body-weight support harness was used during overground walking training during the first 4 weeks of training. In the home setting, however, the harness was anchored to the ceiling and, therefore, was not available for overground walking training. At 2 weeks, the assistive gait device was changed from a rear support walker, which she was never able to successfully control, to a front 4-wheeled walker. A U-Step walker§ was introduced at home 2 months into her home training. As shown in Table 2, there was steady improvement in overground walking, as indicated by a gradual increase in the distance walked from 50 m at 1 week to 152 m at the completion of training, and a reduction in the personal assistance required from maximum assistance from 3 people to supervision by 1 person.


This case report examined the effect of locomotor training using BWST in conjunction with home physical therapy on mobility function in a 13-year-old girl with severe cerebellar ataxia. After 5 months of this training, our patient progressed from being nonambulatory at the start of training to being able to walk 152 m with supervision using a U-Step walker. Other functional improvements included improved independence in transfers. She continued to use her wheelchair for long distances in the community and for independent mobility in the home. Her increased independence in stepping on the treadmill after 4 weeks of training with BWST in the clinic encouraged the family to purchase a BWST system for the home so that training could be continued 5 days a week for 4 months. She continued to make slow, but steady, gains throughout her 5 months of training, despite the fact that training was begun 16 months after her injury.

Previous research28,29 suggests that, following a neurologic injury such as stroke, physical improvements plateau after 6 months. However, recent studies examining the effect of constraint-induced therapy in people with stroke have suggested that gains can be made with intensive therapy several years after the injury.30 Consistent with this research, our patient, whose injury occurred 16 months previously, made significant functional gains with intensive physical therapy (5 days per week for 5 months) that included locomotor training using BWST in conjunction with overground walking.

Our results are similar to those of the case report by Day et al13 on a 9-year-old boy with cerebral palsy who was nonambulatory but was able to complete up to 60 independent steps on the treadmill with body-weight support after 44 sessions of training and demonstrated carryover to overground walking by walking short distances with a rolling walker with minimal assistance 4 months after training. In our patient, functional independence in walking was achieved only after 5 months (approximately 99 training sessions using BWST). This suggests that, in addition to intensity, duration is a critical factor in determining outcomes. Although walking on the treadmill improved steadily over the first month, significant improvements in overground walking were not seen until after 5 months of training.

We believe that a number of factors contributed to the outcomes, including the motivation of the patient and the support and dedication of her family. This family was able to afford a home BWST system and was willing to work with her 5 days per week. In addition, she received in-home physical therapy and occupational therapy several times a week as well as benefiting from additional time from a trained rehabilitation aide. The relative contribution of locomotor training using BWST to the outcomes in light of the fact that she received additional interventions cannot be determined. It is noteworthy, however, that she was nonambulatory after 1 year of in-home therapy, but achieved independent ambulation with the addition of 5 months of intensive locomotor training using BWST.

Results from this case report suggest that locomotor training using BWST is a promising intervention for improving gait in patients with severe cerebellar ataxia who are nonambulatory. Functional gains in walking, however, may require months of consistent practice and training. Findings from this case report provide possible support for research demonstrating the importance of cerebellar structures in locomotor adaptation and in practice-dependent motor learning.16,17 It also supports findings that the rate of locomotor adaptation and thus motor recovery may be slower in the presence of cerebellar pathology compared with other brain regions.16,17 The intensity and duration of locomotor training using BWST required to achieve functional gains in patients with lesser severity of injury is not known. In addition, it is not clear how much additional therapy is needed to achieve these outcomes. Studies are needed to determine the optimal intensity and duration of locomotor training to optimize functional walking outcomes following cerebellar pathology.



Summary of Home Physical Therapy Interventions

aIn-Step Mobility Products Corp, 8027 N Monticello Ave, Skokie, IL 60076.

bNuStep Inc, 5111 Venture Dr, Ste 1, Ann Arbor, MI 48108.


  • Dr Shumway-Cook provided concept/idea/research design. Dr Cernak, Ms Stevens, and Dr Shumway-Cook provided writing, data analysis, and project management. All authors provided data collection. Ms Stevens provided the subject. Mr Price provided facilities/equipment. Mr Price, Ms Stevens, Dr Shumway-Cook, Valerie Kelly, and Stacia Lee, PT, provided consultation (including review of manuscript before submission).

  • * Mobility Research, PO Box 3141, Tempe, AZ 85280.

  • Vision Fitness, 500 South CP Ave, PO Box 280, Lake Mills, WI 53551-0280.

  • Pneumex Inc, 2605 N Boyer Ave, Sandpoint, ID 83864.

  • § In-Step Mobility Products Corp, 8027 N Monticello Ave, Skokie, IL 60076.

  • Received May 3, 2007.
  • Accepted August 7, 2007.


View Abstract