Rehabilitation of balance in two patients with cerebellar dysfunction.

Key Words: Balance rehabilitation, Cerebellar rehabilitation, Postural control.

The cerebellum controls limb, posture, and eye-head coordination and may also be involved in nonmotor functions such as cognition(1) and attention.(2) Anterior lobe (paleocerebellum or spinocerebellum) and midline disease impairs lower-limb coordination, equilibrium responses, and head and trunk synergy, whereas lateral lobe disease chiefly affects limb coordination.(3-5) Anterior and flocculonodular lobe lesions lead to oculomotor and balance impairments, with gait ataxia and functional limitations. Lesions of the flocculonodular lobe result in "central vestibular" symptoms because the peripheral vestibular system can be completely intact, but without cerebellar inhibition, integration of vestibular information is impaired.(3,6) Lesions of the vermis typically produce gait and trunk ataxia, with abnormalities of slow-phase eye movements.(6) Oculomotor impairments caused by cerebellar dysfunction may include saccadic hypermetria, impaired smooth pursuit, increased vestibulo-ocular reflex (VOR) gains, impaired fixation suppression of the VOR, and nystagmus.(3,7)

Functional limitations observed in patients with cerebellar dysfunction may include postural instability,(8,9) gait ataxia,(8) dyssynergia (the inability to perform movement involving multiple joints in one smooth pattern),(10) hypotonicity (decreased resistance to passive stretch, with difficulty fixating limbs posturally),(12) fatiguability, and weakness resulting in activity limitations.(13) The specific mechanism underlying generalized weakness and fatiguability is unclear, but these limitations have been theorized to be due to a loss of cerebellar facilitation to the motor cortex that results in a reduction of spinal motoneuron activity during voluntary movement.(14) The cerebellum also compares sensory information with motor output during voluntary movement and performs predictive compensatory modification of reflexes in preparation for movement. According to Ito,(15) learned movements are either controlled or triggered by processes occurring in the cerebellum. Horak and Diener(16) reported specific deficits in the use of the central set to scale the magnitude of initial postural responses based on prior experience in patients with anterior lobe disorders. Because the cerebellum is considered to be crucial to motor learning, some people believe that patients with known cerebellar dysfunction may be less responsive to physical rehabilitation.

There is limited evidence that treatment programs improve function in patients with cerebellar dysfunction. Recovery after cerebellar lesions or disease in humans is poorly documented. There is, however, strong evidence of recovery after cerebellar lesions in experimental animals, which suggests that if the cerebellum is not totally destroyed, neighboring areas of the cerebellum can adapt or compensate for the impaired region.(9,17) Possible mechanisms of recovery after central nervous system lesions may include neural sprouting, vicarious functions, functional reorganization, substitution, and plasticity.(18)

No studies have demonstrated changes in gait, balance, or locomotor function from exercise interventions for patients with cerebellar dysfunction. Rehabilitation of patients with acute cerebellar dysfunction has included Frenkel's exercises,(19) rhythmic stabilization,(20) and the use of walking aids and weights.(21) Kabat,(20) in 1955, described proprioceptive neuromuscular facilitation (PNF), including resistive exercises to help improve strength, coordination, endurance, balance, and gait, but no research studies of the efficacy of PNF for patients with cerebellar disorders have been reported. There is sparse evidence of successful treatment of chronic cerebellar dysfunction, and some clinicians regard this condition as refractive to treatment.(22) In general, rehabilitation interventions for patients with chronic cerebellar dysfunction have, in the past, been restricted to conservative management (eg, maintaining range of motion) and compensation strategies (eg, recommending that patients increase their base of support or use assistive devices to improve or replace postural stability).(23)

More recently, balance rehabilitation that increasingly challenges body stability have been advocated,(24,25) but most of the available treatment-related publications(24-27) lack adequate intervention descriptions for replication as well as scientific controls. Balliet et al(24) were among the first investigators to describe neuromuscular retraining methods for five patients with chronic cerebellar dysfunction and gait disorders. Treatment was based on the premise that the patients needed to reacquire proper motor control and associated balance through adaptation to increasingly demanding conditions. Therefore, upper-extremity use during balance and gait activities was minimized to facilitate independent balance control. Improvement in functional ambulation was often judged by the type of assistive device used, amount of upper-extremity weight bearing on the device, level of assistance required to ambulate, and maximum distance ambulated. All patients improved on all four variables. Brandt et al(25) proposed similar treatment for ataxia by progressively increasing body instability to activate "sensorimotor rearrangement." In a case report, Sliwa et al(26) reported functional improvement following rehabilitation in a patient with paraneoplastic subacute cerebellar degeneration, but they provided no details about the treatment program.

In our case report of two patients, we describe a staged, home-based intervention approach that provides increasing challenges to body stability in standing and walking. The patients had different etiologies, durations, and clinical presentations of cerebellar dysfunction. The treatment program is based on recent ideas regarding balance rehabilitation suggesting that activities that activate postural and neural control mechanisms may be most effective in achieving better overall postural control.(24-28) Such postural and neural control mechanisms may include the integration of sensory information or alternate motor control strategies to enhance postural stability. We report data regarding patient response to treatment in terms of patient self-report, clinical balance assessment, whole-body movement analysis, and posturography testing (when available).

The medical evaluation for each patient just prior to referral for physical therapy consisted of an examination by a neurologist and diagnostic testing as deemed appropriate by that neurologist. Each patient also underwent a three-dimensional movement analysis in our biomotion laboratory. A whole-body kinematic and kinetic analysis of key standing and gait activities (ie, standing, free and paced gait, and walking in place at a pace of 120 steps per minute) was completed. The motion analysis is described in detail elsewhere.(29-31) Briefly, the system consists of an 11-segment, 66-degree-of-freedom whole-body (head, arms, trunk, pelvis, thighs, shanks, and feet) kinematic model; two force plates; and software to integrate the kinematic and kinetic data.(31) Selspot II hardware(*) and a TRACK kinematic data-analysis software package(+) are used to acquire and analyze the three-dimensional whole-body kinematic data. Floor reaction forces are acquired from two Kistler platforms(++) and processed on the same computer as the kinematic data. Kinematic and kinetic data are sampled at 150 Hz and digitally filtered. The system accuracy is [+ or -]1 mm for linear displacement and [+ or -]1 degree for angular displacement.(32) The accuracy for the estimation of center of gravity (COG) is less than 1 cm.(30)

Stability during the standing and gait tests is quantified in several ways, including 1) using the COG and center of pressure (COP) to calculate the maximum separation that occurs between the COG and COP (or the COG-COP moment arm(30)) during gait tests, 2) standard time-distance variables, such as double-limb support time, speed, and base of support during gait, and 3) phase plane analysis(33) of standing and walking in place. Briefly, a phase-plane analysis involves plotting a variable (eg, mediolateral COG displacement) against its first time derivative (eg, speed).(33) Stable COG standing motion has a converging phase plane, whereas unstable motion has a diverging phase plane (Fig. 1). We use mathematical modeling (root mean square of the plot variance) to quantify the anteroposterior and mediolateral phase-plane plots, thereby permitting quantitative as well as qualitative analysis of the subjects' horizontal COG movement patterns during various activities.(33) Phase-plane analysis has been shown to discriminate between subjects with and without balance impairments.(33) The data collected during the whole-body movement analysis were collected as part of a larger pilot study aimed at identifying the kinetic and kinematic characteristics of postural control that change after rehabilitation. Thus, this information was not used for treatment planning purposes in the following case report of the two patients but is reported for descriptive purposes.

Patient 1--Cerebellar Dysfunction After Removal of a Cerebellar Tumor

History

A 36-year-old left-handed woman was referred for physical therapy. She complained of experiencing dizziness and unsteadiness 7 months following surgical resection of a recurrent pilocystic cerebellar astrocytoma. Six years previous to the surgery, she had severe headaches and hydrocephalus. A cerebellar tumor was identified. Following resection of the tumor, she was back at work within 8 weeks. Five years later, as a result of recurrence of the tumor, the patient had repeat surgery (debulking of residual grade I glioma located in the middle and superior cerebellar vermis and extending to the floor of the fourth ventricle) (Fig. 2). Two months after the resection, she received radiation therapy for 1 month. During radiation treatment, she developed hearing loss, which was worse on the left side. After the second surgery, she noted problems with unsteadiness and dizziness. Her sensation of dizziness worsened following the radiation therapy and was aggravated by head movements. She also reported an unsteady "drunk" sensation. She had moderate dysarthria. A 2-month trial of dexamethasone (4 mg/d), initiated subsequent to radiation therapy to alleviate symptoms of dizziness and unsteadiness, was not effective and was gradually tapered.

The patient lived with her husband. She stopped working as a teacher following the second surgery due to bearing loss, dizziness, and unsteadiness. Due to difficulty with blurred vision during side-to-side head movements, she was not able to drive during the 7 months between surgery and the initiation of physical therapy. She was independent with basic activities of daily living but experienced increased dizziness and unsteadiness in crowded areas such as malls and supermarkets. The patient reported that she ambulated outdoors only when accompanied by her husband because of instability and difficulty crossing streets.

Prerehabilitation Findings

Four days prior to the patient's First appointment for physical therapy, she was examined by a neurologist. The patient told the neurologist that she had resolving left-sided weakness, impaired hand coordination, dizziness, and postural unsteadiness. The neurologist found the patient to be alert and cooperative. Her speech was fluent but mildly dysarthric. Extraocular movements were full, with no spontaneous or gaze nystagmus. She said that her eyes felt "unstable." Facial sensation and movement were symmetrical. Tongue and palate movements were normal. Tympanic membranes were bilaterally intact when inspected visually. Vibration and position sense were present at the toes. Finger-to-nose and heel-to-shin tests were performed accurately and smoothly. Rapid alternating movements were slightly slowed bilaterally. A Romberg's test was negative, but an increased postural sway was observed. Deep tendon reflexes were normal (2+). Gait appeared to be mildly unsteady, with a slightly widened base and a tendency to veer to the left. Step length appeared to be normal, and there was associated arm swing during ambulation. Tandem gait appeared to be unsteady, with a tendency to fall toward either side. Hallpike positional testing revealed no nystagmus and slight dizziness. The patient had impaired fixation suppression of the VOR, which was tested by having her attempt to follow a moving target with the eyes and head moving in unison with the target in the same direction. She demonstrated nystagmus during the test, suggesting that she was unable to suppress the VOR.

Prior to referral for physical therapy, she underwent vestibular testing, including an electronystagmogram (ENG) with caloric stimulation, sinusoidal vertical axis rotation, visual vestibular interaction rotation, and posturography testing using the Equitest system.([sections], 34) During the ENG, there were normal pursuit and saccadic eye movements, no nystagmus with eyes open and closed, and no gaze nystagmus. Sequential closed-loop caloric testing of each ear with 27 degrees C and 44 degrees C water stimulation produced symmetrical and appropriate nystagmus and good fixation suppression of nystagmus. Sinusoidal vertical axis rotation testing with a peak velocity of 50 degrees/s revealed normal VOR gains, phase lead, and symmetry(34) for all seven frequencies tested (ie, 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, and 1.0 Hz). Visual vestibular interaction testing demonstrated normal optokinetic tracking and a normal fixation suppression index. These findings suggested intact peripheral vestibular and brain-stem function. A discrepancy, however, existed between her ENG and rotation test results and her clinical test results with regard to suppression of the VOR. This discrepancy may be explained by the different speeds of head movement at which these various tests are performed. Posturography testing revealed falls on two out of three trials and excessive sway on the single successful trial of standing on a sway-referenced platform with eyes closed (Fig. 3). During the initial physical therapy examination, active range of motion was normal for all extremities. Cervical and lumbar spine mobility was normal and pain-free. Manual muscle test grades were normal (5/5) for all muscle groups of the upper and lower extremities. Finger-to-nose movements and heel-to-shin movements were performed quickly, smoothly, and accurately, bilaterally. Rapid alternating movements were slow on the left side. Sensation for light touch and proprioception (as tested by noting the patient's ability to accurately describe the joint position at the knees, ankles, and toes) was normal. Muscle tone was normal. She reported slight dizziness during positional testing (moving to a side-lying position to either side from the sitting position), but there was no nystagmus. Extraocular movements were normal. She was unable to visually fixate her gaze on a target when performing side-to-side head movements at frequencies greater than 1 Hz (timed with a metronome). With her head stationary, visual acuity was 20/10. With side-to-side head movements at 1 Hz, visual acuity worsened to 20/40, and was further degraded to 20/70 when head movements were performed at 2 Hz, which suggested to us that she had difficulty with VOR-mediated gaze fixation at these speeds of head movement.

Sensations of the frequency and intensity of dizziness were evaluated using a standardized questionnaire. The patient reported that dizziness was present for some part of every day at a 5/10 intensity (10 being the highest level of dizziness imaginable). Dizziness increased with head movements in a sitting or standing position and with attempts to focus on objects when walking. The sensation of disequilibrium was present daily at a 7/10 intensity. The patient scored 88 out of a possible 100 on the Dizziness Handicap Inventory (DHI),(35) reporting problems in eight out of nine items related to functional activities, in all seven items related to physical activities, and in eight out of the nine items related to emotional health (Tab. 1). The DHI, originally devised to measure perception of handicap in individuals with benign paroxysmal positional vertigo, is used routinely in our clinic for patients with dizziness or balance problems to document perception of handicap related to the dizziness or balance problems.

For both patients, balance during standing and walking was evaluated using a standardized protocol that has been in place at our clinic for 5 years. Both patients were evaluated and treated by the same physical therapist. Balance assessment focused on evaluating each patient's sensory and motor organization aspects of postural control as well as automatic postural responses. These tests were selected based on the patient's neuropathology and history, and the need to identify (for treatment planning purposes) the underlying aspects of postural control that were most problematic for each patient.

In standing and walking, the patient had a tendency to be positioned posteriorly (ie, weight appeared to be shifted over the heels). She could stand with feet together and eyes closed for 60 seconds (measured with a digital stopwatch). When she was asked to maintain balance in that position and turn her head from side to side at a pace of 60 beats per minute (paced with an audible metronome), balance could be maintained for 60 seconds with a marked increase in postural sway. She was able to maintain standing balance on a 7.62-cm (3-in) compliant cushion with eyes open for 30 seconds and with eyes closed for 3.4 seconds, suggesting that she had difficulty with postural control when visual inputs were removed and somatosensory inputs were altered.

She was able to maintain balance in tandem standing with eyes open for 60 seconds. Standing on the dominant leg (determined by first asking the patient which leg would be used to kick a ball) could be performed for 30 seconds with eyes open and for 2.74 seconds with eyes closed, suggesting difficulty with postural control when visual inputs were removed and she needed to maintain her COG over a narrow base of support. Increased postural sway in the posterior direction was observed during all standing balance tests, especially when the base of support was narrowed or when visual cues were removed. This finding suggested to us that she may have had difficulty selecting appropriate sensory input for postural control. Stepping strategies (taking a step to regain balance) were delayed in response to unexpected perturbation or release in the anteroposterior direction. The timed "up and go" test(36) (performed over a 12.2-m [40-ft] walkway) was completed in 14.34 seconds without the use of an assistive device. The patient's gait appeared to be steady, and she slowed down during a 180-degree turn. Walking with eyes closed over a 6.1-m (20-ft) walkway was performed in 5.6 seconds, but with three crossed steps (one foot crossed over the other or a stagger to the side), suggesting difficulty with postural control during gait when visual inputs were removed. She was able to perform tandem gait with eyes open for only two steps but was unable to perform this task with eyes closed, suggesting difficulty with postural control during gait when the base of support was narrowed and the COG needed to be maintained within a smaller area. Walking over a 6.1-m walkway turning her head from side to side every third step was performed in 7.02 seconds, but with two crossed steps, suggesting difficulty integrating inputs related to head movements with other sensory inputs during this activity.

The patient's goals for rehabilitation included being able to drive and walk outdoors independently. She also hoped to decrease her sensation of dizziness during head movements so that she could become comfortable in crowded areas and be able to participate more in social activities.

Rehabilitation

The balance rehabilitation program designed for this patient was based on the following interpretation of the patient's condition:

1. Her impaired postural stability was related to cerebellar dysfunction resulting from pathology of the midline cerebellum. Even though she did fairly well on most of the balance tests, she exhibited increased postural sway during the tests, which could reflect difficulties with selecting appropriate sensory information for postural control. The practice of increasingly demanding balance and gait activities may help improve the patient's ability to select and use visual, somatosensory, and vestibular inputs more effectively to minimize her unsteadiness. Improved steadiness may decrease her perception of disequilibrium and improve her ability to control the COG.

2. She was most dependent on visual and somatosensory cues for her standing balance, and she demonstrated an inconsistent ability to use vestibular information for postural control in situations in which visual and somatosensory information were less available than vestibular inputs. For example, she was unable to stand on foam with eyes closed for more than a few seconds and, during posturography testing, had difficulty standing on a sway-referenced platform with eyes closed (Fig. 3), a testing situation in which vestibular inputs are primarily available for postural control. She was able, however, to stand on a sway-referenced platform with a sway-referenced visual surround (although she demonstrated excessive sway on the initial trial), a testing situation in which visual inputs are thought to be critical to resolve the imposed sensory conflict. Her vestibular function tests showed normal peripheral vestibular function. The vestibular system is modulated centrally by the cerebellum.(6) The cerebellum pathology, therefore, is the likely cause of this patient's inability to consistently use vestibular inputs effectively.

3. Her delay with stepping strategies (taking a step to regain balance) during perturbation testing (quickly pushing the patient's COG outside the base of support) in the clinic may be consistent with the scaling problems (matching the magnitude of the response to the displacement) seen in patients with cerebellar dysfunction.(16) For example, her delay in stepping backward with posterior perturbations may be due to an initial hypermetric (or exaggerated) response to the perturbation(14) (which presumably would occur in the tibialis anterior muscles), causing her COG to move anteriorly. The antagonist muscles may then contract (in a hypermetric manner as well),(16) causing a posterior displacement of her COG and perhaps leading to taking a step. The repetition of balance and gait activities may help her to better utilize somatosensory feedback to scale (or match) the magnitude of these postural responses.(16)

4. Her impaired gaze stabilization during functional activities was related to her cerebellar dysfunction. Because patients with cerebellar dysfunction can use vision to help achieve postural stability,(37) we believe that it is important to try to improve gaze stability to improve stability during gait and other activities of daily living that typically require head movement. Her impaired gaze stability may be related to either her VOR function or her inability to suppress her VOR. Therefore, activities that help the patient use alternative strategies for gaze stabilization (eg, using cervical inputs during head movements via the cervico-ocular reflex) as well as practice suppressing the VOR may be helpful.

Our balance rehabilitation treatment program for this patient consisted of the standing balance and gait exercises and activities outlined under phase 1 through phase 3 in Table 2. She performed phase 1 of the program during the first 2 weeks, phase 2 during weeks 3 and 4, and phase 3 during weeks 5 and 6. In addition, she performed eye-head coordination exercises to help improve gaze stabilization during side-to-side head movements (Appendix, page 552). A brief explanation of the rationale for each activity in the treatment program is included in Table 2. The patient attended weekly 30- to 45-minute physical therapy sessions, during which her status was reassessed, she participated in eye-head coordination exercises as well as gait and balance activities, and she was instructed in progression of her home program. She had no difficulty with weekly progression of her home exercise program. She was instructed to perform the exercises at least once daily at home and to note the exercises completed, number of repetitions, and any difficulties experienced using a standard adherence tool in use at our clinic. She completed daily exercise logs, indicating that she was adherent to the home program on a once-per-day basis 5 out of 7 days per week over the 6-week course of treatment. The home program took her 30 to 40 minutes to complete.

After 6 weeks of treatment, she was reevaluated in the clinic and in the biomotion laboratory using the same measures as those used during the initial assessment. Repeat ENG and posturography tests were also performed.

Postrehabilitation Findings

At the conclusion of the 6-week treatment program, this patient reported a decrease in frequency and intensity of disequilibrium. She now felt unsteady only occasionally (ie, up to three times per week) at an intensity of 5/10 (compared with 7/10 prior to rehabilitation). There was no change in the frequency or intensity of dizziness. She scored 72/100 on the DHI (compared with 88/i00 at the initial visit), reporting improvements in two physical activities, in two functional activities, and in three areas of emotional health that were reported to be problematic at the initial visit (Tab. 1). No items on the DHI were reported to be worse. The patient was now able to drive and walk outdoors independently. Balance assessment revealed no change in the ability to perform standing on foam with eyes closed or unilateral standing with eyes closed. Performance on other standing balance measures also remained the same, but there was a consistent decrease in the amount of postural sway observed. There was no change in the timed "up and go" test. Walking with eyes closed was performed in 5.39 seconds, but with three crossed steps (slightly faster). Tandem gait with eyes open could be performed for three steps (compared with two steps initially); she was still unable to perform tandem gait with eyes closed. No delay in response was observed during unexpected perturbation or release in the anteroposterior direction. There was an improvement in her ability to stabilize gaze during head rotation, as indicated by less degradation of visual acuity with head rotation at 1 Hz (20/20 compared with 20/40 initially) and at 2 Hz (visual acuity was 20/30 as compared with 20/70 initially).

Repeat ENG, sinusoidal axis rotation, and visual vestibular interaction testing showed no change when compared with testing done prior to rehabilitation. Repeat posturography testing indicated that she was able to maintain balance with normal equilibrium scores on all six test conditions (Fig. 4), thereby showing an improvement in her ability to stand on a sway-referenced platform with eyes closed.

The three-dimensional movement analysis revealed several changes indicative of improved postural stability (Tab. 3). During preferred pace (or free) gait, speed increased while base of support decreased, indicating that the patient moved more quickly and with a narrower base of support during gait. During paced gait (ie, when the speed of gait was controlled), double-limb support time and base of support decreased, implying that the patient walked in a more stable manner. The whole-body maximum COG-COP moment arm increased (whereas it decreased in the mediolateral direction) during both free and paced gait, implying a higher level of stability.(30,38) The COG phase-plane measure during semitandem standing improved (Fig. 1), indicating that the patient had better control of her COG displacement and speed during these rusks.(33,39)

Patient 2--Cerebellar Dysfunction Due to Cerebrotendinous Xanthomatosis

History

A 48-year-old right-handed man was referred for physical therapy with a diagnosis of cerebrotendinous xanthomatosis (CTX) resulting in balance and gait instability. Cerebrotendinous xanthomatosis is a rare autosomal recessive disorder of lipid metabolism (deficiency of liver mitochondrial enzyme required for oxidation of cholesterol to bile acids), characterized by elevations in cholestanol and cholesterol that deposit in tendons, peripheral tissues, in the spinal cord, and centrally in the cerebellum.(40) Major clinical manifestations include progressive cerebellar ataxia, subnormal intelligence, tendon xanthomas, cataracts, dementia, and limb paresis due to spinal cord pathology.(40) The onset of this progressive neurodegenerative disease typically occurs in early adulthood.(40)

This patient initially noted swelling of the Achilles tendon and gait difficulty at age 38 years. His gait instability progressed over the 10 years prior to our seeing him and resulted in a few falls during the last 2 years. His medical history included a T-12 compression fracture 4 years previously, a fracture of the left lower extremity as a result of a fall 2 years previously, and bilateral lens implants 5 years ago for cataracts. The patient had been maintained on cholesterol-lowering medication (Chenodil, 250 mg, three times daily) for the past 4 years. He had been seeing a neurologist twice a year to monitor his metabolic and neurologic status.

Transcranial Doppler testing provided evidence of occluded left vertebral artery and basilar artery disease. An electroencephalogram indicated mild generalized slowing of background activity and occasional brief episodes of more pronounced slowing over the left frontal temporal region. Sterol analysis demonstrated consistently increased concentrations of cholestanol and other sterols (however, the values were greatly reduced from values recorded 5 years previously). A computed tomography scan of the head performed approximately 3 years prior to referral for physical therapy showed cerebellar atrophy, calcification in the dentate nuclei of the cerebellum, and widening of the prepontine region consistent with brain-stem atrophy (Fig. 5). Trace to mild mitral regurgitation, with prolapse of both mitral leaflets, was shown on an echocardiogram. Carotid ultrasounds were normal. Electromyography showed no definitive electrophysiologic evidence of sensorimotor neuropathies. He lived with his mother and a younger brother, who also had CTX. He stopped driving a few years ago due to progressive visual problems related to his cataracts. He was employed as a materials handler in an industrial company but stopped working 2 years prior to his referral for physical therapy, at which time he applied for disability due to his visual problems and increasing difficulty with standing balance and walking.

Prerehabilitation Findings

A neurologist who examined him 2 months prior to his referral for physical therapy found him to be pleasant, talkative, and oriented to person, place, and time. There was mild dysarthria and decreased hearing on the right side. Extraocular movements and pursuit tracking were normal in all directions. There was a mild lateral gaze nystagmus on far excursion bilaterally. Lipid deposits measuring approximately 8 X 10 cm were present on both Achilles tendons. Deep tendon reflexes were brisk (3+) bilaterally in the upper extremities, normal (2+) at the knees, and diminished (1+) at the ankles. Babinski testing was normal bilaterally. Motor examination (performed with resisted isometric muscle testing of all muscle groups) revealed normal and symmetric upper- and lower-extremity strength. There was mild dysmetria in the left upper extremity. There was no evidence of dystonia or choreoathetosis. Light touch, temperature, and proprioception were intact in all limbs. Vibration sense was mildly decreased in the toes and ankles bilaterally. The patient was able to walk independently without an assistive device, but his gait was wide-based (approximately 51 cm [20 in]) and appeared to be unsteady and stiff. Turns during gait were made very slowly and without rotation of the trunk. He was unable to perform tandem gait. Previous neuropsychological testing indicated that he was mildly retarded, without language or aphasic difficulties. The patient's neurologic examination findings were unchanged compared with the findings of an examination done 6 months earlier.

During the initial physical therapy examination, the patient was able to follow two-step commands but was easily distracted. Active range of motion in the upper and lower limbs was normal except for ankle dorsiflexion, which was limited to 0 degrees bilaterally (with the knee extended). Manual muscle testing of the upper and lower extremities was normal (5/5) for all muscle groups. Heel-to-shin movements were slow but accurate bilaterally. Finger-to-nose and rapid alternating movements of the upper extremity were slow and slightly dysmetric on the left side. The patient's sensation of light touch and proprioception were intact in both lower extremities.

The patient did not report any sensation of dizziness (eg, feelings of spinning or lightheadedness). He reported the sensation of disequilibrium (eg, as if he might fall) approximately two or three times per week (when standing or walking) at an intensity of 6/10. A DHI was not completed due to the patient's cognitive limitations and his inability to distinguish problems related to his dizziness and balance from his visual problems. He was able to go outdoors independently, and he routinely spent 2 to g hours each day visiting friends. He reported feeling the need to be cautious when he was outside alone due to his gait instability and as a result decreased his walking speed and distance. He did not routinely help with household tasks but was able to help with cleaning and laundry. He found going outdoors to be particularly difficult during the winter months. As a result, he restricted walking outdoors in the winter and would go out only if he had transportation.

The patient demonstrated a forward-bent posture during standing and walking. He stood with a 20.3-cm (8-in) base of support. He said that he felt that he was steadier in this position as compared with when his feet were closer together. He was able to stand erect in response to verbal cues; however, he reported feeling unsteady and experienced the sensation of falling backward. The patient could stand with feet together and eyes closed for 19 seconds. When he was asked to maintain balance in that position and turn his head from side to side at a pace of 60 beats per minute, balance could be maintained only for 7 seconds. This finding suggested to us that the patient had difficulty when inputs related to head movement needed to be integrated simultaneously, He was able to maintain standing balance on a 7.62-cm compliant cushion with eyes open for 30 seconds and with eyes closed for 11 seconds, suggesting difficulty with postural control when vision was removed and somatosensory inputs were altered. He was unable to perform tandem standing with eyes open, suggesting difficulty maintaining the COG over a narrow base of support. He could stand on one leg for 1.5 seconds with eyes open but was unable to do so with eyes closed. This finding suggested to us that the patient had difficulty maintaining the COG over a narrow base of support even when visual inputs were available. A marked postural sway in the anteroposterior direction was observed during all standing balance tests, especially when the base of support was narrowed.

The timed "up and go" test (performed over a 12.2-m walkway) was completed in 18 seconds without the use of an assistive device. His gait appeared to be steady, although he slowed down during the 180-degree turn. Walking with eyes closed over a 6.1-m walkway was performed in 10 seconds, but with two crossed steps, suggesting difficulty with postural control during gait when vision was removed. He was unable to perform tandem gait with eyes open or closed. This finding suggested to us that the patient had difficulty maintaining the COG over a narrow base of support during gait even when visual inputs were available. Walking over a 6.1-m walkway while turning the head from side to side every third step was performed in 9 seconds, but with two crossed steps, suggesting that the patient had difficulty integrating vestibular inputs from head movements during walking. Gait on a 2.54-cm (1-in) foam walkway was steady and performed in 4 seconds with eyes open. Stepping strategies were markedly delayed in response to a sudden perturbation (a quick, high-speed push on the sternum to move the COG outside the base of support) or release (an unexpected removal of pressure from the sternum) in the posterior direction.

The patient's goals for rehabilitation included being able to walk outdoors for 4.8 km (3 miles) with a steadier gait, improved posture, and decreased risk and fear of falling.

Rehabilitotion

The balance rehabilitation program designed for this patient was based on the following interpretation of his condition:

1. His primary problem of impaired postural stability was related to his cerebellar dysfunction. The pathology causing his cerebellar dysfunction was presumed to involve diffuse regions of his cerebellum rather than a specific area. Patients with cerebellar degeneration tend to have increased anteroposterior sway in standing with eyes open or closed when compared with individuals without cerebellar dysfunction, but they are able to use vision to decrease their unsteadiness.(37) This patient's postural sway did not appear to decrease even when his eyes were open during clinical balance testing (eg, standing with feet together or on foam), suggesting that he may not be using vision effectively to help decrease his unsteadiness. He may have difficulty using vision effectively for postural control due to his impaired visual acuity related to his cataracts. This patient may benefit from balance retraining that focuses on improving his use of visual cues for postural control.

2. He was not effectively using vestibular information for postural control in situations in which visual and somatosensory information were less available than vestibular inputs, as noted by his performance during balance testing (eg, standing on foam with eyes closed, walking with eyes closed). This finding was presumed to reflect abnormal integration of sensory information in the cerebellum and to be related to his cerebellar dysfunction.

3. His difficulty with stepping strategies, supported by his performance during clinical testing with perturbation tests, may be consistent with scaling problems seen in patients with cerebellar dysfunction.(16) Alternately, it could be due to his inability to maintain balance (ie, control his COG) while standing on one leg. Repetition of specific balance and gait activities may help him to better utilize somatosensory feedback for scaling the magnitude of his postural responses.

4. His decreased ankle range of motion was related to decreased flexibility of his plantar-flexor muscles bilaterally and could contribute to his impaired postural responses (eg, inability to perform tandem standing or one-legged standing). Active stretching exercises, therefore, were included in the treatment program.

The balance rehabilitation treatment program received by the patient consisted of the standing balance and gait activities outlined under phase 1 through phase 3 in Table 4. The patient performed phase 1 of the program during the first 2 weeks, phase 2 during weeks 3 and 4, and phase 3 during weeks 5 and 6. A brief explanation of the rationale for each activity in the treatment program is included in Table 4. In addition, he was given exercises to stretch the ankle plantar flexors to help improve or preserve ankle range of motion. The patient had no difficulty tolerating progression of the exercises during weekly 30-minute physical therapy sessions. He was instructed to perform the exercises at least once daily at home and to note the exercises completed, number of repetitions, and any difficulties experienced using a standard adherence tool in use at our clinic. Due to evidence of slight cognitive impairments, his mother supervised his assigned home exercise program, which took 30 minutes to complete. The patient completed daily exercise logs, indicating that he was adherent to the home program 4 out of 7 days per week on a once-per-day basis over the 6-week course of treatment. After 6 weeks of treatment, he was reevaluated in the clinic and in the biomotion laboratory using the same measures as those used during the initial assessment.

Postrehabilitation Findings

At the conclusion of the 6-week treatment program, the patient reported a decrease in both the frequency and intensity of his disequilibrium. He reported the sensation of disequilibrium only occasionally (approximately one time per week), and the intensity was reduced to 3/10 (from 6/10 initially). He was able to walk for 4.8 km several times per week, with less fear of falling. No falls had occurred. There was no change in ankle range of motion. Balance assessment revealed improvements in his ability to stand with feet together with eyes closed (60 seconds compared with 19 seconds) and with eyes closed with head rotation (17 seconds compared with 7 seconds). He improved in his ability to maintain balance with eyes closed on foam (21 seconds compared with 11 seconds). Tandem standing with eyes open could be performed for 5 seconds (as compared with 0 seconds); standing on one leg with eyes open could be performed for 3 seconds (as compared with 1.5 seconds). There was no change in his ability to stand on one leg with eyes closed. The timed "up and go" test was performed in 14 seconds (19% faster). Gait with eyes closed was performed 11% faster, with only one crossed step. The patient was able to perform tandem gait for three steps (compared to no steps before) with eyes open but was still unable to do tandem gait with eyes closed. Walking with head rotation was performed 19% faster, with no crossed steps (compared with two steps before). Walking on a foam walkway was performed 43% faster. Stepping strategies now appeared to be only minimally delayed in response to sudden perturbation or release in the posterior direction.

Several changes indicative of improved postural stability were shown with he three-dimensional movement analysis (Tab. 3). During free gait, walking speed increased while base of support decreased, indicating that the patient walked faster with a more narrow base of support. During paced gait, double-limb support time and base of support decreased, further suggesting that the patient walked in a more stable manner. The whole-body maximum COG-COP moment arm increased (whereas it decreased in the mediolateral direction) during free and paced gait, implying a higher level of stability.(30,38)

Discussion

Both patients demonstrated improvements in postural stability and function while reporting improvements in their perception of disequilibrium following 6-week individually designed programs of physical therapy. Although there were some similarities in the patients' responses to treatment, there were important differences between these two patients. This discussion will address two important questions: 1) Why did the patients have different responses to treatment, as indicated by varied performance results on test measures? and 2) If the treatment programs were related to the improvements, why might the improvements have occurred?

The patient with the recurrent resected cerebellar tumor (initially diagnosed 6 years prior to the initiation of physical therapy) reported problems with dizziness and unsteadiness only 7 months prior to starting physical therapy. Therefore, her problems were relatively recent in onset, and she was in the subacute phase of recovery. In contrast, the patient with CTX reported progressively worsening problems with unsteadiness over 10 years. The patient with the resected cerebellar tumor may, therefore, have a greater potential for residual spontaneous recovery even though her symptoms were stable during the 7 months prior to referral for physical therapy.

Another important consideration is the difference in neuropathology between the two patients. In the patient with cerebellar tumor resection, the middle and superior vermis was predominantly affected (Fig. 2) and her functional problems were predominantly related to gait ataxia, consistent with a lesion in the vermis. In the patient with CTX, there was evidence of diffuse cerebellar atrophy, suggesting involvement of the spinocerebellum (both the vermis and intermediate zones of the hemispheres), the cerebrocerebellum, and the vestibulocerebellum (Fig. 5). In addition, there was CT scan evidence of involvement of the dentate nuclei (Fig. 5). Lesions of the deep cerebellar nuclei (such as the dentate nucleus) typically produce symptoms that are more severe than those seen with disorders restricted to the cortex.(41) This patient's clinical signs and symptoms, however, suggested greater involvement of the spinocerebellum. This patient did not have formal testing (ie, vestibular function tests) to examine central and peripheral vestibular pathways; therefore, involvement of the vestibulocerebellum cannot be ruled out.

Why might postural stability have improved in both of these patients? Bronstein et al(37) showed that although patients with cerebellar dysfunction swayed more than individuals without cerebellar dysfunction during tests with eyes open, with eyes closed, and with visual stimulation, they retain the ability to use vision to control much of their unsteadiness. Furthermore, these investigators found that on repeatedly moving the visual surround, the patients were able to suppress the destabilizing effect of visual Stimuli, demonstrating the ability to adapt their responses and switch reliance to vestibulo-proprioceptive-postural loops.(37)

The patient with the resected cerebellar tumor did fairly well on balance measures at initial evaluation and therefore demonstrated minimal posttreatment changes on timed balance measures. This finding may have occurred because the patient underwent adaptation or spontaneous recovery prior to referral for physical therapy. Other considerations are that this patient was younger than the patient with CTX and that she may have already learned to use vision to control much of her unsteadiness. Her posttreatment posturography findings demonstrate her improved ability to integrate and use appropriate vestibular information in the absence of visual and somatosensory information. Why she did not improve in her ability to maintain balance while standing with feet together and eyes closed on compliant foam is unclear. This is a clinical test that is designed to simulate (to some degree) the same condition on which she improved during posturography testing. Perhaps standing on a foam surface serves me)re to distort than to reduce somatosensory input (such as occurs during posturography testing), making it a more difficult task for this patient. In contrast to the patient with CTX, this patient showed marked improvement in her ability to control COG speed and displacement in semitandem standing (Fig. 1). A decreased postural sway was also evident in the anteroposterior direction during assessment of the same standing posture. These findings indicate that time-based measures alone may not be adequate to accurately detect changes in postural stability in a patient such as this one, whose initial assessment revealed minimal deficits in performance.

The patient with CTX showed marked improvement in his ability to maintain balance with eyes closed on a firm surface both with and without head movements. This finding implies that the patient may have become better trained to use somatosensory inputs for postural control. He also demonstrated improved postural stability on a compliant foam surface with eyes closed, demonstrating his improved ability to use vestibular inputs. Similarly, his performance improved during tandem gait with eyes open and walking with head rotation, which suggests that he learned to use visual inputs more effectively for postural control. One possibility is that the patient, through repetition of the various activities performed, was trained to more efficiently use any available sensory input (visual, somatosensory, or vestibular). Many studies now support the idea that cerebellar circuits are modified by experience and that these changes are important for motor learning.(41) This patient, however, did not improve on COG phase-plane measures during standing with feet together and eyes open or closed, or in semitandem standing with eyes open. That the COG phase-plane measures did not improve after physical therapy quantifies the clinical observation of increased postural sway during these standing tasks. A more prolonged period of practice (eg, longer than 6 weeks) might have resulted in improvements in this patient's ability to control the speed and displacement of his COG, but further study is needed to determine whether this is true.

Although these two patients demonstrated different changes in standing postural stability after treatment, both patients demonstrated similar trends in improvements on gait stability measures (Tab. 3), despite their initial differences in performance and abilities. Briefly, free gait speed increased, whole-body COG-COP maximum moment arm increased (whereas COG-COP maximum moment arm in a mediolateral direction decreased), and the base of support decreased. When the speed of gait was controlled, the improvements demonstrated in base of support, double-limb support time, and mediolateral COG-COP maximum moment arm Were of similar magnitude for the two patients.

Horak and Deiner(16) have suggested that the anterior lobe of the cerebellum in humans is not critical for normal latencies or for spatial or temporal timing of multisegmental agonists and antagonists of automatically triggered postural responses. The anterior lobe, however, appears to be critical for accurately tuning tile magnitude (or scaling) of postural responses based on immediate prior experience. Use of speed feedback to scale the magnitude of postural responses is not impaired in patients such as those studied by Horak and Diener, who had various types of cerebellar dysfunction but all involving some portion of the anterior lobe.(16) Horak and Diener concluded that patients with anterior cerebellar lobe dysfunction appear to compensate for the abnormal scaling of hypermetric agonist responses by using somatosensory feedback to adjust the magnitude of later-occurring antagonist responses to the speed and amplitude of actual postural displacement. Both patients described in our case report had some involvement of their anterior cerebellar lobes, although neither patient had exclusive anterior lobe pathology.

The physical therapy program for both patients included activities in standing and during gait that would challenge their postural stability (eg, standing or walking with a narrow base of support with eyes open and closed), thereby inducing a movement strategy in response (eg, multisegmental sway, ankle sway). These experiences were probably instrumental in improving both patients' ability to effectively use somatosensory feedback for modifying the magnitude of postural responses rather than relying on prior experience or central set. This theory might also explain why both patients did better on the perturbation tests. An unexpected perturbation in the posterior direction, for example, produces an initial posterior perturbation, and the patient then has to return the COG to a more central point to stay erect or take a step if the COG has moved outside the base of support. Patients with cerebellar dysfunction have hypermetric responses to unexpected perturbations, causing them to sway more and for a longer time before returning to an equilibrium point (Bortolami SB, Wernick M, Krebs DE; unpublished research). After treatment, both patients appeared to have more timely stepping strategies when challenged with a large unexpected perturbation, suggesting that they were better able to scale the magnitude of their postural responses. Further research, including electromyographic analysis of muscle activity during perturbation tests, is necessary to determine whether this is true.

Summary

After participation in individually designed, staged programs of physical therapy, two patients with different cerebellar pathologies demonstrated improvements in symptoms, postural stability, and function, as indicated by self-report and three-dimensional movement analysis of gait activities. One patient did not demonstrate changes on time-based clinical tests but showed improvements on kinematic and kinetic variables measured during three-dimensional movement analysis of standing tasks. This finding indicates that time-based measures may not be comprehensive in detecting improvement in postural stability for all patients. We believe that it is unlikely that improvements in performance noted could be attributed to practice of the tasks, because the three-dimensional movement analysis of the gait activities included activities that were not performed as part of the treatment program (ie, paced controlled gait). Whether the trends in improvement described for these two patients using whole-body movement analysis represent trends that characterize other patients with similar pathologies cannot be determined. These conclusions can only be made after studying a large group of patients in a more controlled manner. The similarities in the two patients' responses to treatment emphasize that patients with cerebellar lesions, acute or chronic, can learn to improve their postural stability. Therefore, if the cerebellum is not totally destroyed, adaptation or compensation for the impairments produced appears to occur, perhaps in a neighboring area of the cerebellum or in another part of the brain. Further research is necessary to investigate the effectiveness of this treatment approach for patients with cerebellar dysfunction and to further understand the strategies used by persons with balance impairments to improve their postural control.

(*) Selective Electronics Corp, Partille, Sweden.

(+) Developed at the Massachusetts Institute of Technology, Cambridge, MA 02139.

(++) Type 9281A, Kistler Instruments AG, Winterthur, Switzerland.

[sections] NeuroCom International Inc, 9570 Lawnfield Rd, Clackamas, OR 97015.

Table 1.
Dizziness Handicap Inventory Items(35) Reported to Be
Problematic Prior to Rehabilitation and Improvements
Reported After Rehabilitation for Patient With Resected
Cerebellar Tumor

Physical/
Functional/
Emotional        Sometimes a              Always a
Factors          Problem                  Problem

Physical         Performing household     Looking up(a)
activities       chores                   Walking down a
                 Turning over in bed      supermarket aisle(b)
                 Walking down a           Quick head movements
                 sidewalk                 Bending over

Functional       Getting into or out      Traveling
activities       of bed                   Social activities(a)
                                          Reading(a)
                                          Managing heights
                                          Strenuous housework
                                          Walking outdoors
                                          independently(b)
                                          Walking in the dark
                                          (at home)

Emotional                                 Frustration
factors                                   Fearful to go out
                                          alone(a)
                                          Fearful of being home
                                          alone(a)
                                          Embarrassment(a)
                                          Concentration
                                          Feelings of being
                                          handicapped
                                          Depression
                                          Increased stress on
                                          relationships

(a) Item rated as "a little improved" after rehabilitation.
(b) Item rated as "much improved" after rehabilitation.

Table 3.
Percentage of Improvement From Pretreatment Evaluation in
Kinematic Indicators of Stability" During Standing Balance
Activities and Locomotor Performance for Two Patients With
Cerebellar Dysfunction

                              Improvement         Improvement
                                (%)b                (%)b
                             Demonstrated        Demonstrated
                             by Patient With    by Patient With
                             Cerebellar           Cerebellar
Task/Variable                 Tumor              Xanthomatosis

Free gait
Speed                           20                    10
Base of support                 36                    24
Whole-body COG-COP
 maximum moment arm(c)           4                    16
Mediolateral COG-COP
 maximum moment arm(d)          18                    18
Paced gait
 Cycle time                      0                    10
 Double support time            28                    28
 Stance duration                 2                    18
 Base of support                36                    29
 Whole-body COG-COP
  maximum moment arm            43                     3
 Mediolateral COG-COP
  maximum moment arm            24                    24
Anteroposterior COG
 phase plane(e)
 Feet together with
  eyes closed                   24                   -54(f)
 Semitandem stance
  with eyes open                70                No data(g)

a COG=center of gravity, COP= center of pressure.

b Improvement corresponds to a decrease in the values for all kinematic indicators of stability except speed and whole-body COG-COP maximum moment arm; for these items, improvement reflects an increase in these values.

c Whole-body COG-COP maximum moment arm is the difference between tire body's COP and COG in the diagonal (combined anteroposterior and mediolateral) direction. A larger whole-body moment arm indicates that a state of less biomechanical stability is allowed to occur during the activity; that the patient allows the overall moment arm to get larger during an activity (and does not fall) signifies a higher level of overall balance control.(30) A 10% improvement in whole-body COG-COP maximum moment arm has been shown 10 represent a statistically significant and functionally meaningful difference in individuals with balance impairments.(38)

d Mediolateral COG-COP maximum moment arm is the difference between the body's COP and COG in the mediolateral direction. A smaller maximum moment arm in the mediolateral direction during activities in which the body's COG and COP are moving anteriorly (eg, walking forward) indicates that the individual is more stable and walking with a less variable base of support.

e Anteroposterior COG phase plane is a measure of the variance of the COG speed and displacement in the anterioposterior direction. A smaller value indicates less variance in the COG speed and displacement during a task, suggesting a higher overall level of stability.(33) Based on data from asymptomatic subjects, a change of more than 30% represents a meaningful change.(39)

f This value represents a decline in stability.

g No data available because the patient was unable to complete the test.

[TABULAR DATA OMITTED]

References

1 Roland PE. Partition of the human cerebellum in sensory-motor activities, learning, and cognition. Can J Neurol Sci. 1993;20(suppl 3):S75. Abstract.

2 Akshoomoff NA, Courchesne E. A new role for the cerebellum: cognitive operations. Behav Neurosci. 1992;106:731.

3 Gilman S, Bloedal J, Lechtenberg R. The symptoms and signs of cerebellar disease. In: Disorders of the Cerebellum. Philadelphia, Pa: FA Davis Co; 1981:189-262. Contemporary Neurology series.

4 Dichgans J, Diener HC. Different forms of postural ataxia in patients with cerebellar disease. In: Igarishi M, Black FO, eds. Disorders of Posture and Gait. New York, NY: Elsevier Science Inc; 1986:207-215.

5 Adams RD, Victor M. Abnormalities of movement and posture due to disease of the extrapyramidal motor systems. In: Principles of Neurology, Part II: Cardinal Manifestations of Neurologic Disease. New York, NY: McGraw-Hill Inc; 1989:68-75.

6 Thurston SE, Leigh RJ, Abel LA, Dell'Osso LF. Hyperactive vestibuloocular reflex in cerebellar degeneration: pathogeneses and treatment. Neurology. 1987;37:53-57.

7 Zee DS, Yee RD, Cogan DG, et al. Ocular motor abnormalities in hereditary cerebellar ataxia. Brain. 1976;99:207-234.

8 Cooper IS. Involuntary Movement Disorders. New York, NY: Harper & Row, Publishers Inc; 1969.

9 Amici R, et al. Cerebellar Tumor Resections: Clinical Analysis and Physiopathologic Correlations. New York, NY: S Karger and Co; 1976:35-76.

10 Holmes G. The cerebellum of man. Brain. 1939;62:1.

11 Oas JG, Baloh JW. Vertigo and the anterior inferior cerebellar artery syndrome. Neurology. 1992;42:2274-2279.

12 Iloeje SO. Measurement of muscle tone in children with cerebellar ataxia. East Afr Med J. 1994;71:256-260.

13 Viallet F, Bonnefoi-Kyriacou B, Massion J, et al. Quantitative assessment of postural asynergy in cerebellar pathology. In: Trouillas P, Fuxe K, eds. Serotonin: The Cerebellum and Ataxia. New York, NY: Raven Press Inc; 1993:343-355.

14 Bremer F. Le Cervelet. In: Roger GH, Binet L, eds. Traite de physiologie normale et pathologique, volume 10. Paris, France: Masson; 1935.

15 Ito M. Neurophysiologic aspects of the cerebellar motor control system. Int J Neurol. 1970;1:162.

16 Horak FB, Diener HC. Cerebellar control of postural scaling and central set in stance. J Neurophysiol. 1994;72:479-493.

17 Ito M. The Cerebellum and Neural Control. New York, NY: Raven Press Inc; 1984:1-7.

18 Bach-y-Rita P. Central nervous system lesions: sprouting and unmasking in rehabilitation. Arch Phys Med Rehabil. 1981;62:413-417.

19 Urbsheit NL, Oremland BS. Cerebellar dysfunction. In: Umphred DA, ed. Neurological Rehabilitation. 2nd ed. St Louis, Mo: CV Mosby Co; 1990:597-618.

20 Kabat H. Analysis and therapy of cerebellar ataxia and asynergia. Arch Neurol Psychiatry. 1955174:375-382.

21 Morgan MH. Ataxia and weights. Physiotherapy. 1975;61:332-334.

22 Sage GH. Motor Learning and Control: A Neuropsychological Approach. Dubuque, Iowa: Wm C Brown Communications Inc; 1984.

23 Morgan MH. Ataxia: its causes, measurement, and management. Int Rehabil Med. 1980;2:126-132.

24 Balliet R, Harbst KB, Kim D, Stewart RV. Retraining of functional gait through the reduction of upper extremity weight-bearing in chronic cerebellar ataxia. Int Rehabil Med. 1987;8:148-153.

25 Brandt T, Krafczyks S, Mahbenden I. Postural imbalance with head extension: improvement by training as a model for ataxia therapy. Ann NY Acad Sci. 1981;74:636-649.

26 Sliwa JA, Thatcher S, Tet J. Paraneoplastic subacute cerebellar degeneration: functional improvement and role of rehabilitation. Arch Phys Med Rehabil. 1994;75:355-357.

27 Safe AF, Cooper S, Windsor ACM. Cerebellar ataxia in the elderly. JR Soc Med. 1992;85:449-451.

28 Shumway-Cook A, Woolacott MH. Assessment and treatment of patients with postural disorders. In: Shumway-Cook A, Woolacott MH, eds. Motor Control: Theory and Applications. Baltimore, Md: Williams & Wilkins; 1995:207-235.

29 Krebs DE, Lockert J. Vestibulopathy and gait. In: Spivack BS, ed. Evaluation and Management of Gait Disorders. New York, NY: Marcel Dekker Inc; 1995:93-116.

30 Riley PO, Hodge WA, Mann RW. Modelling the biomechanics of posture and balance. J Biomech. 1990;23:503-5-05.

31 Krebs DE, Wong DK, Jevsevar DS, et al. Trunk kinematics during locomotor activities. Phys Ther. 1992;72:505-514.

32 Antonsson EK, Mann RW. Automatic 6-DOF kinematic trajectory acquisition and analysis. Journal of Dynamic Systems Measurement and Control. 1989;111:31-39.

33 Riley PO, Benda BJ, Gill-Body KM, Krebs DE. Phase plane analysis of stability in quiet standing. J Rehabil Res Dev. 1995;32:227-235.

34 Parker SW. Vestibular evaluation: electronystagmography, rotational testing, and posturography. Clin Electroencephalogr. 1993;24:151-159.

35 Jacobson GP, Newman CW. The development of the dizziness handicap inventory. Arch Otolaryngol Head Neck Surg. 1990;116:424-427.

36 Podsiadlo D, Richardson S. The timed "Up and Go": a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39: 142.

37 Bronstein AM, Hood JD, Gresty MA, Panage C. Visual control of balance in cerebellar and parkinsonian syndromes. Brain. 1990;113: 767-779.

38 Krebs DE, Gill-Body KM, Riley PO, Parker SW. Double-blind placebo-controlled trial of vestibular rehabilitation for bilateral vestibular hypofunction: preliminary report. Otolaryngol Head Neck Surg. 1993;109:735-741.

39 Danis CM. Relationship of Standing Posture and Stability. Boston, Mass: MGH Institute of Health Professions; 1996. Master's thesis.

40 Brown MS, Goldstein JL. Disorders of lipid metabolism. In: Petersdorf RG, Adams RD, Braunwald E, et al, eds. Harrison's Principles of Internal Medicine. 10th ed. New York, NY: McGraw-Hill Book Co; 1983:547-559.

41 Ghez C, Fahn S. The cerebellum. In: Kandel E, Schwartz J, eds. Principles of Neural Science. 2nd ed. New York, NY: Elsevier Science Inc; 1985:502-522.

KM Gill-Body, PT, NCS, is Assistant Professor, Graduate Programs in Physical Therapy, MGH Institute of Health Professions, 101 Merrimac St, Boston, MA 02114 (USA) ([e-mail]), and Neurologic Clinical Specialist, Physical Therapy Services, Massachusetts General Hospital, Boston, Mass. Address all correspondence to Ms Gill-Body at the first address.

RA Popat, PT, NCS, was Supervisor, Physical Therapy Services, Massachusetts General Hospital, and Adjunct Assistant Professor, MGH Institute of Health Professions, at the time this report was written. She is currently a graduate student, University of Massachusetts, Amherst, Mass.

SW Parker, MD, is Chief of Otoneurology, Massachusetts General Hospital, and Assistant Professor of Neurology, Harvard Medical School, Boston, Mass.

DE Krebs, PhD, PT, is Associate Professor, MGH Institute of Health Professions; Director, Massachusetts General Hospital Biomotion Laboratory, Boston, Mass.; Instructor, Harvard Medical School; and Lecturer, Massachusetts Institute of Technology, Cambridge, Mass.

This work was supported by National Institute of Disability and Rehabilitation Research grant HI 33G60045 and National Institutes of Health grant RO1AG11255.