An observational study of balance and proprioception function in patients with spinocerebellar ataxias type 3


 BackgroundPostural instability is one of the most disabling features of spinocerebellar ataxias type 3 (SCA3) and often leads to falls that reduce mobility and functional capacity. This study aimed to quantitatively analyse static and dynamic balance and proprioception function on postural control in patients with SCA3 using the Pro-kin system and optimise rehabilitation programmes for them.MethodsEight-one clinically diagnosed SCA3 patients (38 women, 43 men; aged 39.00 ± 9.66) and 62 healthy controls were studied and evaluated using the Pro-kin system (PK254P, Tecnobody S.r.l, Dalmine, Italy). The measurements included (1) a static balance test in two visual feedback conditions: eyes open (EO) and eyes closed (EC); (2) a dynamic balance test measuring limits of stability (LOS); and (3) a proprioception function test to obtain proprioceptive measurements on a multiaxial balance evaluator for both right and left lower limbs.ResultsCompared to controls, SCA3 patients showed significantly higher values of all static balance outcome variables with eyes open and eyes closed, implying postural instability. SCA3 patients showed significantly higher values in the standard deviation of body sway along the medio-lateral (ML) axis and in the velocity of body sway along the anterior-posterior (AP) axis. The overall scores and the scores for all eight LOS components were significantly lower in the SCA3 patients than in the controls. The mean values of AP index (API), ML index (MLI), Stability index (SI) and average trace error (ATE) were significantly greater in SCA3 patients compared to HC subjects, while API showed a trend toward higher values.ConclusionsSCA3 patients have a significant postural control disorder, and are likely to fall on the AP plane and prefer performing postural adjustments in the ML direction; a decreased proprioception function in the knee and ankle is also evident. Visual cues and proprioception should be emphasized in balance rehabilitation training. Attention should also be paid to improve muscle strength and range of motion.Trial registrationThe Chinese clinical test registration center. ChiCTR1800020133. Registered 15 december 2018 - Retrospectively registered, http://www.chictr.org.cn/showprojen.aspx?proj=33950

An observational study of balance and proprioception function in patients with spinocerebellar ataxias type 3 Spinocerebellar ataxias (SCAs) are a large, complex group of autosomal dominant neurodegenerative diseases characterised by progressive cerebellar ataxia with oculomotor dysfunction, dysarthria, pyramidal and extrapyramidal signs, peripheral neuropathy and cognitive impairments [1][2] . To date, more than 40 types of SCAs are described [3] . Spinocerebellar ataxia type 3 (SCA3), which is also known as Machado-Joseph disease (MJD), is currently the most common subtype of SCAs, accounting for nearly 50% of all SCA patients in the Chinese population [4] . SCA3 is caused by a cytosine-adenineguanine (CAG) trinucleotide repeat expansion on chromosome 14q32.1, resulting in an expanded polyglutamine repeat in the encoded ataxin-3 protein [5][6] .
Balance impairments in SCA3 are characterised by increased postural sway and poor balance control during both static and dynamic tasks [7] . These impairments are the most frequent initial clinical manifestation in SCA3, which can be quickly recognised by patients as a sign of onset [8] . As the disease progresses, SCA3 patients usually show an uncoordinated, unstable, wide-based gait [9] . Postural stability is one of the main factors of SCA3 directly affecting gait [7][8] . Various underlying complications, such as muscle atrophy, change of muscle tone and joint motion limitation, give rise to postural instability [10] .
Such de cits in SCA3 patients are progressive in nature and are strongly associated with an increased risk of falling [11][12] , which in turn affects their posture and balance control. These problems are considered the main cause of worsening the quality of life and independence of these individuals [13] .
Studies have shown that rehabilitation training may improve the balance function of SCA patients [14][15][16] , but there is no consistent recommendation for rehabilitation training. It is therefore necessary to evaluate postural stability to optimise rehabilitation programmes for SCA3 patients to improve their function, reduce the risk of falling and improve their quality of life.
In a study that assessed balance in patients with SCA, Aizawa et al. [17] used the Tinetti balance test, the Scale for the Assessment and Rating of Ataxia (SARA) and the International Cooperative Ataxia Rating Scale (ICARS) to evaluate disease severity. A study used the Berg Balance Scale (BBS) to observe the balance in patients with SCA [18] . In the EuroSCA fall study [11] , patients were asked to complete a fall questionnaire and the SARA. However, these common semi-quantitative clinical scales and questionnaires for balance assessment used in these studies are, to some extent, subjectively in uenced [19] and cannot detect minor balance de cits in mildly disabled patients [20] . Further, these scales are often used in the later stages of the disorder, meaning early identi cation of patients who have a potential for postural instability is not possible [21] . The development of an objective method for identifying postural stability in patients with SCA3 in the early stages is therefore essential.
Computerised posturography tests have the advantage of being more sensitive and speci c than clinical tests. Bunn L M et al. [22] used an infrared motion-capture system to investigate stance instability in SCA6 and determine how it is affected by varying stance width. Nanetti L et al. [23] used the Tetrax® posturography system to show a progressive impairment of stance performance in the SCA1 preclinical phase. In a study, the analysis of trunk motion and muscle responses in SCA patients was shown using the SwayStar system [24] . Nonetheless, subjects of prior studies almost always examine other SCA subtypes, and most posturography studies have focused mainly on measuring body sway using the stability index. Very few studies have attempted to determine the characteristics of postural stability from both static and dynamic aspects, which provide a foundation for characterizing patients' postural and balance instability [25] . In addition, the proprioception function has been shown to yield additional information on postural stability [26] , yet to the best of our knowledge, there is no study on this aspect in SCA3 patients. The Pro-kin system is a new type of visual feedback instrument equipped with a balance force platform and computer [27] . This system can be used to test static and dynamic balance and the proprioception function and overcomes the de ciencies of balance detection in prior studies.
We previously used the Pro-kin system to assess the static and dynamic stability of SCA3 clinically, and the results proved to be effective [28] . To the best of our knowledge, our previous study is the only study about static and dynamic balance in SCA3 patients. However, no intensive study exists that examines the overall postural stability in SCA3 patients. Our study therefore to analyze the differences between SCA3 patients and healthy control people in both balance ability and proprioception function using the Pro-kin system; our purpose is to nd the targets requiring key intervention, so as to guide the selection and design of clinical rehabilitation training methods and improve the effect of rehabilitation training intervention.

Participants
A cohort of 81 patients (43M/38F, aged 39.00 ± 9.66) with a diagnosis of molecular-con rmed SCA3 [29] was assessed at the Department of Rehabilitation  lack of sensitivity in the lower limbs; (5) musculoskeletal, cardiovascular or respiratory system impairments or other accompanying ailments; (6) engagement in another rehabilitative study protocol; or (7) recent oral medication that would affect balance.
A cohort of 62 age-and gender-matched healthy control (HC) subjects (31M/31F, aged 40.18 ± 10.99) were enrolled as a reference population. The controls showed no evidence of balance de cits; comorbidities, such as diabetes, hypertension, rheumatic, or oncological diseases; vision problems; or musculoskeletal, cardiovascular, or respiratory system problems. The age, height and weight of the subjects were collected.
The protocol was approved by the Ethics Committee of the First A liated Hospital, Fujian Medical University. The study's design and procedures were performed in accordance with the Declaration of Helsinki. Written informed consent was obtained prior to participation.

Experimental Design
We conducted an observational study between SCA3 patients and HC subjects at the same time of the day. The Pro-kin system (Prokin 254 (Pro-Kin Software Stability), TecnoBody S. r. l., Dalmine, 24044 Bergamo, Italy) was used to assess postural stability, which is based on assessing the movement of the centre of pressure (COP) and measuring proprioceptive on a multiaxial balance evaluator for lower limbs. A rehabilitation therapist blinded to the groups conducted the assessments. All participants received the following postural stability assessments in a naturally and brightly lit and quiet room: (1)

Balance Measurement
We placed and xed the four locks of the system under the force platform to conduct the static balance test [27] . All participants stood barefoot on the platform in a standard standing position with arms hanging comfortably at their sides and feet placed symmetrically at the corresponding position on the platform. In the EO trials, participants were instructed to look straight ahead and gaze at a speci c target 80 cm away.
In the EC trials, the participants were denied visual feedback, but they were instructed to face forward as if looking straight ahead. Each trial lasted 30 seconds. Five key parameters are measured [30] : (1) [31] . COP is the point of application of forces exchanged between feet and ground. SD is de ned as the mean error of the COP displacement, EA is the ellipse that contains the COP trajectory, and SP is the length that contains the COP trajectory. Speci cally, the highest SD value identi es the preferred direction of the postural adjustments performed by the subject. The highest velocity value indicates that the sway is signi cant in this direction. The Romberg of the EA (R EA ) and SP (R SP ) were evaluated as the ratio between EC and EO. Higher R EA and R SP values re ect greater instability with closed eyes. In total, we measured 14 outcome variables based on ve parameters (Fig. 2).
The circular force platform can tilt 20° in all directions from the horizontal, and the degree of surface instability of the platform can be adjusted from level 10 (most stable) to level 1 (least stable) using the microprocessor-based actuator control incorporated in the system. The most stable level (level 10) was used to examine the dynamic balance test, and the medium stable level (level 5) was used in the proprioception function test in all participants.
In the dynamic balance test, all participants were asked to stand on the platform to perform the LOS test.
They were asked to shift their centre of mass, without changing their foot position, towards the targets that appeared randomly in eight different directions only once. The direction of the target was indicated on the screen by a blinking target. Participants were required to achieve the target from the centre position using the shortest vertical or horizontal path. The path used was given a score by the instrument;  [32] . In a particular direction, the maximum achievable perfect score was 100. A lower score indicated greater sway (Fig. 3).
For the proprioception function test, participants placed the right foot on the balance board and the left foot on a xed support platform equal to the balance board [33] . Participants had to focus on the monitor and draw lines on the screen by moving their right foot (on the balance board) within 120 seconds. Their motor task is to try to superimpose the lines drawn by the movements of their foot on those drawn by the system. Then, the left lower limb was tested for two key parameters: (a) balance control indices: AP index (API), ML index (MLI) and stability index (SI) [°] and (b) the Average Trace Error (ATE) [%] [33] , the preassessment of lower limb proprioception. The smaller the value, the better the lower limb proprioception function and balance control (Fig. 4).
To reduce the risk of falls and minimise interference from external support, a trainer stayed close alongside or behind the participants.

Statistical Analyses
All statistical analyses were performed using GraphPad version 8.0.1. The normality of the distribution of all postural parameters variables was assessed using Kolmogorov-Smirnov tests. Following the results of these tests, variables with normal distribution and variables in non-normal distribution were expressed as mean ± SD and median (range), respectively. Two-independent sample t-tests and Mann-Whitney U tests were used to analyse normally and non-normally distributed measures, respectively, and to con rm the statistical difference of the two groups and the two lower limbs according to the normality assumption. Chi-squared tests were used to assess the gender difference between SCA3 patients and HC subjects.
Statistical test outcomes were considered signi cant at p < 0.05.

Results
Static Balance Test Analysis Table 1 shows the values of all static balance outcome variables for the two groups. We compared the values between SCA3 patients and HC subjects for the EO and EC conditions separately. The statistical analysis revealed that all indices were signi cantly worse for SCA3 patients compared to HC subjects (all ps < 0.0001), suggesting worse balance in SCA3 patients. Particularly, in both EO and EC conditions, the SD of body sway values along the ML axis were all greater than the AP axis in SCA3 patients and velocity of body sway values along the AP axis showed a trend toward higher values than the ML axis. Variables were represented as the mean ± standard deviation and for variables with skewed distribution, the median was presented in parentheses Abbreviations: EA,ellipse area; SP,sway path; SDap and SDml, standard deviation of COP displacement in the anterior-posterior and medio-lateral direction;Vap and Vml,velocity of body sway along anterior-posterior and medio-lateral direction. 1 Mann-Whitney U test We then quanti ed the impact of loss of visual feedback on balance performance by comparing the Romberg indices (R EA and R SP ) (Fig. 5) Table 2 shows the LOS scores of the SCA3 patients and HC subjects. The SCA3 patients had signi cantly lower total LOS score and overall scores for all eight components of the LOS scores compared to the HC subjects, the difference between groups was signi cant (all ps < 0.05). In SCA3 patients, the LOS score for all directions was asymmetrically affected. In particular, the forward LOS scores was lower than other components of the LOS scores.

Discussion
Although previous studies have examined postural control problems in spinocerebellar ataxias patients, the nature of postural instability and the pathophysiological and biomechanical mechanism of balance in SCA3 patients remain unknown.Also, no targeted rehabilitation program has been proposed in these studies. The Pro-kin system is an ideal tool for the evaluation of balance function at present. It can not only judge the cause and degree of balance function damage, but also evaluate the effect of treatment and rehabilitation. In this study, the Pro-kin system was used to assess the static balance, dynamic balance and proprioception function of SCA3 patients. Our main ndings show that the function of visual afference affects postural control in SCA3 patients; these patients have predominant instability in the AP plane and prefer performing ML direction postural adjustments; the distribution of centre of gravity in SCA3 patients is asymmetrical, and they have a worse ability to shift the weight forward; notably, SCA3 patients have a decreased proprioception function, mainly in the knee and ankle joints.
The SCA3 patients consistently exhibited increased postural instability in all experimental conditions compared with HC subjects. We observed the existence of a predominant alteration of body sway velocity in the AP axis in SCA3 patients in both EO and EC conditions. This demonstrated that our patients have more AP falls. Mohan et al. [34] quantitatively assessed the balance in spinocerebellar ataxia type 1 and found that SCA1 patients had global impairment of balance and a signi cantly greater body sway in the AP direction than in the ML direction. A previous study of trunk movements revealed that autosomal dominant spinocerebellar ataxia patients have worse trunk sway and predominant instability in the AP direction [24] . Our study is in line with these related studies. Therefore, maintaining good control of body sway in the AP direction should be of high priority in preventing falls and in balance training programs in patients with SCA3.
Besides these similar phenomena, our study also found that SCA3 patients showed a signi cant trend toward higher values in the standard deviation of body sway along the ML axis. The ndings suggest that SCA3 patients prefer performing ML direction postural adjustments. Thus, the ML standard deviation is one of the most reliable markers of postural instability in SCA3 [35][36] and can be used to analyse postural instability in SCA3 in future studies. Moreover, this nding can explain why SCA3 patients adopt an abnormally typically broad-based gait and have marked di culties performing tandem gait [9] . The increased values in the ML standard deviation may re ect an attempt to maintain stabilising movements during a quiet stance, which may be a compensatory strategy to reduce their intrinsic instability in the AP direction. Impairments in the activation function of the synaptic transmission between the climbing bres and Purkinje cells inhibits cortical motor activation via a complex neural pathway involving the dentate nucleus, which could be related to abnormal postural sways in SCA3 patients [37][38] . Previous studies have demonstrated that cerebellar transcranial magnetic stimulation (TMS) is capable of facilitating motor cortical activation via modulation of Purkinje cell excitability [39][40] . Therefore, TMS is recommended to activate the function of the cerebellar to improve balance.
The main functions of the cerebellum are to maintain postural stability, regulate muscle tone and coordinate the voluntary movement of muscles. Control of postural stability is a multifaceted process and involves the integration of sensory information from proprioception, vision and vestibular systems [41] . As one of the main results, the presented ndings highlight that SCA3 patients had statistically signi cant higher Romberg indices values in both R EA and R SP compared to HC subjects, which re ected that an absence of visual control or insu cient input of visual information inenhances an increase in postural sway in SCA3 patients [24] . During the dynamic balance process, asymmetrically affected component LOS scores in all eight directions indicated that SCA3 patients have less adaptive capacity to effect correct postural control in all directions according to the task requirement, even with visual cues [34] . Our results contradict the common physiological model in which vision helps control postural stability [42] . Owing to the neuronal loss occurrings in the basal ganglia, SCA3 patients have a de cit in reweighting various sensor-motor loops. When adapting to novel situations, this de cit affects the integration of sensory information for postural stability [43] . This may re ect the predominant involvement of the spinocerebellum (anterior lobe) in SCA3, as the anterior lobe is associated with visual input [44] . Therefore, visual cues may be required in balance rehabilitation so as to compensate for the decreased balance function caused by cerebellar factors in SCA3.
The results of this study showed that the total LOS score and overall scores for all eight components of the LOS scores of patients with SCA3 were smaller, which was signi cantly different from that of healthy controls. Also, LOS score of forward was the lowest and LOS score of back was highest in all eight components of the LOS scores in SA3 patients, which was consistent with the results measured in the health control group under the same conditions. In normal activities, the range of body stability limit is smaller than the theory, and the range of stability limit is tilted at different angles in multiple directions.
There are more activities in forward and back direction in human activities, so the limit of stability range in forward and back direction has greater in uence on daily life [45] . Melzer et al. [46] thought that the forward LOS was related to the muscle strength of exor metatarsus and extensor dorsum of the ankle joint, and believed that the muscle strength of metatarsophalangeal exor played a more obvious role in the prevention of falls. It is suggested to increase the imitative movement exercises such as retrieving in the balance training, and increase the forward and backward movement range through the muscle strength training of metatarsal exus, so as to contribute to the body balance.Therefore, in the balance training of patients withSCA3, attention should be paid to the training of muscle strength of trunk and lower limbs, so as to expand the range of stability limits of forward and backward, especially forward.
Proprioception is a nerve impulse that is sent to the central nervous system by mechanical receptors located in joints, joint capsules, ligaments, muscles, tendons and skin. It can be divided into strength sense, motion sense and position sense. Instead of weakening somatosensory feedback by standing on foam to analyse the interactions between postural stability and proprioceptive function [47] , in this work, we performed the lower limb proprioceptive function test. To the best of our knowledge, this study is the rst to analyse the proprioception of the left and right feet separately in SCA3 patients. We found that our patients obtained larger values for API, MLI ,SI and ATE compared to HC subjects, and there was a trend toward higher values in API, suggesting that postural instability in SCA3 patients correlates with a de cient proprioception function and a quantitative reduction in muscle strength, mainly in the knee and ankle joints [26] . The pathological involvement of spinocerebellar proprioceptive input and the loss of the integrity of the medial somatosensory descending system may explain abnormal postural control [48] .
Patients presented a locking of knees and ankles and muscular rigidity, causing abnormal joint movements related to postural instability. When human body is about to fall after receiving small and slow interference, the body mainly relies on ankle joint regulation to restore postural stability (ankle joint strategy). The decline of ankle joint position sense affects the implementation of ankle strategy, which may be the cause of poor balance in SCA3 patients. We suggest that SCA3 patients receive a proprioceptive-motor training rehabilitation program and stretching and strength exercises. Unexpectedly, no signi cantly better proprioception function was observed for the right lower limb compared to the left lower limb in our SCA3 patients. Since all the patients are right dominant, it remains to be seen why patients' right lower limbs lost the advantage of a better proprioception function to control balance.
As our study was an observational study, we have not provided information on changes in postural stability over time in our patients. As posturography cannot identify the speci c constraints underlying postural instability, the predictive validity of these measurements in monitoring disease progression remains un-explored.

Conclusions
In this study, the Pro-kin system was used to analyse static and dynamic balance and proprioception function on postural control in patients with SCA3 using the Pro-kin system. Our patients have a signi cant postural control disorder; they are likely to fall in the AP plane and prefer performing ML direction postural adjustments. Decreased proprioception function in the knee and ankle was also observed. Impaired postural stability in our SCA3 patients relates to abnormal integration of the somatosensory descending system and/or inappropriate cerebellar motor commands. The gure shows the Pro-kin system.(a) patient was tested balance function (b)patient was tested in the proprioception function  The gure shows the actual recording of the path of postural sway during Limits of stability (LOS) task in (a) a control subject and (b) a SCA3 patient.

Figure 4
The gure shows the actual recording of the path of postural sway during Proprioceptive multiaxial assessment in right lower limb in (a) a control subject and (b) a SCA3 patient. Figure 5