This study aimed to identify quantitative biomarkers of motor function for cerebellar ataxia by evaluating gait and postural control using an RGB-depth camera-based motion analysis system. In 28 patients with degenerative cerebellar ataxia and 33 age- and sex-matched healthy controls, motor tasks (short-distance walk, closed feet stance, and stepping in place) were selected from a previously reported protocol, and scanned using Kinect V2 and customized software. The Clinical Assessment Scale for the Assessment and Rating of Ataxia (SARA) was also evaluated. Compared with the normal control group, the cerebellar ataxia group had slower gait speed and shorter step lengths, increased step width and mediolateral trunk sway in the walk test (all P < 0.001). Lateral sway increased in the stance test in the ataxia group (P < 0.001). When stepping in place, the ataxia group showed higher arrhythmicity of stepping and increased stance time (P < 0.001). In the correlation analyses, the ataxia group showed a positive correlation between the total SARA score and arrhythmicity of stepping in place (r = 0.587, P = 0.001). SARA total score (r = 0.561, P = 0.002) and gait subscore (ρ = 0.556, P = 0.002) correlated with mediolateral truncal sway during walking. These results suggest that the RGB-depth camera-based motion analyses on mediolateral truncal sway during walking and arrhythmicity of stepping in place are useful digital motor biomarkers for the assessment of cerebellar ataxia, and could be utilized in future clinical trials.
Research Article
Digital motor biomarkers of cerebellar ataxia using an RGB-depth camera-based motion analysis system
https://doi.org/10.21203/rs.3.rs-3089838/v1
This work is licensed under a CC BY 4.0 License
Journal Publication
published 18 Sep, 2023
You are reading this latest preprint version
Cerebellar ataxia
gait ataxia
postural control
motion analysis
RGB-depth sensor
Kinect
Ataxia is a health condition in which the muscles fail to function in a coordinated manner, resulting in an inability to maintain posture or perform movements or actions smoothly. It should be noted that ataxia is a distinct condition from motor paralysis. Specifically, patients with ataxia have poor movement accuracy, impaired coordination between agonistic and antagonistic muscles, and increased movement variability. Ataxia may be sensory, cerebellar, or both in origin. Sensory ataxia resulting from large fiber sensory neuropathy and/or dorsal column impairment presents as reduced proprioception, reduced deep tendon reflexes, and Rombergism [1]. Cerebellar ataxia (CA) is a general term used to describe motor symptoms induced by cerebellar dysfunction. CA predominantly affects aspects of movement, including the coordination of balance, gait, speech, and limb and eye movements. Of these, gait and balance ataxia are considered crucial factors affecting daily activities and may increase the risk of falls [2–4]. Ataxic gait is characterized by decreased speed and step length along with increased step width and variability in the timing and direction of steps [5]. Multi-joint coordination patterns across the hip, knee, and ankle are abnormal in patients with CA [6]. Features of postural dysfunction in CA include increased body sway and base width during standing or walking, which may be related to impaired predictive postural adjustments and disrupted coordination between the head, trunk, and legs, known as asynergia [6]. There exists a relationship between the specific body part affected by ataxia and the site of the cerebellar lesion. Ataxia of the limbs is linked to lesions of the anterior cerebellar hemispheres, while lesions in the flocculonodular lobe are associated with vestibular control of dynamic stability during walking and eye-head coordination. Additionally, truncal ataxia is associated with lesions in the cerebellar vermis [7]. The causes of CA are acquired and genetic; the former include neurodegenerative, vascular, autoimmune, demyelinating, toxic, infectious, and neoplastic types, while the latter include autosomal dominant, autosomal recessive, X-linked, and mitochondrial types [8]. The most common autosomal dominant CA types in Japan are spinocerebellar ataxia type 3 (SCA3), spinocerebellar ataxia type 6 (SCA6), dentato-rubullo-pallido-luysian atrophy, and spinocerebellar ataxia type 31 (SCA31) [9, 10]. Among sporadic adult onset degenerative CA, multiple systemic atrophy (MSA) is characterized by CA and involvement of autonomic function and extrapyramidal signs.
Quantifiable markers of motor function are increasingly being explored as useful tools to aid in the clinical ratings of neurodegenerative diseases. Several testing protocols and technologies have been applied to assess CA in different body parts and are often evaluated using clinical ratings [11, 12]. Among the clinical ratings, the Scale of the Assessment and Rating of Ataxia (SARA) semi-quantitatively describes the severity of CA with reasonable sensitivity and reproducibility, both cross-sectional and longitudinal assessments [13, 14]. To conduct multicenter longitudinal therapeutic trials, there is a need to test biomarkers that allow for more objective, eligible detection of subtle changes and feasible quantitative measures. Recently, applied technologies have included wearable sensors [2, 15, 16] and three-dimensional marker-free cameras [3, 17–19] for quantifying movement patterns in neurological diseases. Motognosis Labs (Motognosis, https://www.motognosis.com/en/technology/index.html) is a marker-free motion analysis application that uses a commercially available RGB-depth sensor (Microsoft Kinect V2; Microsoft, Redmond, USA), accompanied with customized analytical software to derive functional motor parameters from tasks, including walking, balance, and postural control while standing [17]. The motor tasks used in this system were developed to evaluate motor function in neurological diseases with high reproducibility [17] and have previously been applied and validated in healthy individuals [17, 20, 21], patients with multiple sclerosis [22, 23], and patients with Parkinson's disease [24].
This study aimed to elucidate the motor characteristics of CA using an RGB-depth camera-based motion analysis system, the Motognosis System, with multiple motor tasks. The analyses consisted of two steps. First, quantifiable motor parameters of gait and postural control were generated using the Motognosis system; these were compared between CA patients and healthy individuals to describe the motor dysfunction of CA, thereby defining potential diagnostic biomarkers. Second, the relationships between these motor parameters and clinical ratings of disease severity (SARA scores) were investigated through correlation analyses to explore the clinical relevance and potential sensitivity. These results will help identify the most promising tasks and quantitative motor parameters that are applicable for diagnosis and longitudinal observations in future studies as digital motor biomarkers of CA.
Participants
From July 2018 to July 2021, 31 CA patients who were able to walk without aids were enrolled in this study, including patients with SCA3 (n=3), SCA6 (n=8), SCA31 (n=3), dominantly inherited CA with unknown genetic classification (n=2) and one with sporadic adult onset CA (n=1), or MSA with predominant cerebellar ataxia (MSA-C; n=14). All autosomal dominant SCAs were diagnosed based on genetic testing results. Sporadic CA was diagnosed as progressive ataxia with onset after the age of 20 years, with no family history or known cause of ataxia [8, 25]. MSA-C fulfilled the ‘probable’ diagnostic criteria of the second consensus statement [26]. None of the participants exhibited any space-occupying lesions or coarse ischemia on brain MRI scans. Three SCA3 patients were excluded from the analyses because of concomitant involvement of the sensory nervous system, leaving 28 patients with CA. None of these patients displayed clinical features of spasticity, weakness, sensory deficits, or cognitive impairment. Clinical assessment of ataxia in patients with CA was performed using SARA [13]. The total SARA score ranges from 0 to 40, with higher scores indicating more severe symptoms. Item scores of gait (item 1, score 0–8), stance (item 2, score 0–6), sitting (item 3, score 0–4), and heel-shin slide (item 8, score 0–4), were used as subscores for correlation analyses.
All participants were able to walk independently without aid. Further, 33 age- and sex-matched healthy controls were included if there was no history of trauma that would result in motor dysfunction and no cognitive dysfunction. This study was approved by the Ethics Committee of the Graduate School of Medicine, Chiba University (approval no. 3174). Written informed consent was obtained from all the participants.
Motor function assessment
All motion data were acquired using an RGB-depth camera-based motion analysis system (Motognosis Labs, Version 1.2.0; Motognosis GmbH, Berlin, Germany) using a real-time 3D consumer camera (Microsoft Kinect V2) for motion capture. The recorded movement information included depth data streams and a three-dimensional time series of 25 artificial anatomic landmarks in RGB video streams using the Kinect SDK V2.0.1400.
The motor tasks for the assessment consisted of 4 m of short comfortable speed walking (SCSW), posturography with open eyes (POCO) for 20 s, and stepping in place (SIP) for 40 s (Table 1). All tests were set up, instructed, and measured by the same examiner according to the standard operating procedures of the Motognosis System. The patients were either placed 2.5 m (SIP and POCO) or 5 m (SCSW) from the camera system in a standing position, as specified for each test (Fig. 1). Owing to the technical specifications of the sensor, the walking length was limited to approximately 4 m [17]. The SCSW task was recorded three times per assessment, whereas POCO and SIP were only recorded once. The recorded data were used to extract 12 spatiotemporal parameters describing different aspects of motor performance. The SCSW parameters are reported as the average of all three replicates. Descriptions of each motor parameter are provided in Table 2.
The formulae for calculating the arrhythmicity (step time coefficient of variation) (1) and asymmetry (2) in the SIP are based on the general definitions described below [27, 28]
Statistical analysis
The normality of the parameters was explored through inspection of group-wise histograms and Shapiro–Wilk tests. For demographic values, sex differences were compared using the Chi-square test, and continuous data were compared using the unpaired t test or Mann-Whitney U test, appropriately. Mean, standard deviation (SD), and coefficient of variation (CV) were calculated for all the derived parameters in each group. Pearson’s correlation or Spearman’s correlation was used to explore the influence of age, body height, and weight on motor parameters in HC, as appropriate. The Bonferroni-correction was performed to compare measurements at the group level. Correlation analysis between each motor parameter and SARA score in the CA group was performed using Pearson’s correlation or Spearman’s correlation analysis, as appropriate. Bonferroni correction was performed and a P-value of less than 0.005 was regarded as statistically significant. Calculations were performed using IBM SPSS statistics 28.0.1.1 ((IBM Corporation, Armonk, NY, USA). Diagrams were created using the seaborn-package version 0.7.1 and Matplotlib-package version 2.0.0.
Relationship between motor parameters and demographic values
All demographic and motor parameter data were normally distributed, except SARA subscales, SIP asymmetry, SIP stance time. Summarized clinical features of the patients with CA and healthy individuals are shown in Table 2. Total SARA score in the CA group was 11.8 (2.9) [mean (SD)]. There were no differences in body height, body weight, body mass index (BMI), age, or sex distribution between patients and controls (P > 0.19). Demographic data between 14 MSA-C and 14 other CAs were not different, except for disease duration, which was shorter in MSA-C (MSA-C: 1.4 [1.0] years, other CAs: 7.9 [4.6] years, mean [SD]).
In healthy individuals, age was negatively correlated with gait speed (ρ = -0.517, P = 0.002), step length (ρ = -0.688, P < 0.001), and positively correlated with anterior-posterior trunk deflection range during standing (ρ = 0.589, P < 0.001). Body height was negatively correlated with trunk deflection range in the anterior-posterior during standing (ρ = -0.508, P = 0.003), and positively correlated with step length (r = 0.566, P < 0.001). There was no correlation with BMI. The results of motion analysis in healthy individuals revealed no significant differences between the sexes in any motor parameter when corrected for height and age.
Group comparisons of motor parameters
A comparison of the 12 motor parameters between the healthy control group and patients is summarized in Table 3. In the SCSW task, compared to the healthy controls, patients with CA showed a reduction in step length (P < 0.001), a decrease in gait speed (P < 0.001), and an increase in step width (P < 0.001). Cadence did not differ between the CA and control groups (P = 0.16). Trunk deflection range during walking showed an increase in the mediolateral direction (P < 0.001) but not in the anterior-posterior direction (P = 0.19) in the CA group compared with that in the control group. In the POCO task, the trunk deflection range was larger in the mediolateral direction in the CA group than in the control group (P < 0.001). In the SIP task, the CA group demonstrated an increase in arrhythmicity (P < 0.001) and a prolonged stance time (P < 0.001) compared to the healthy control group.
Relationship between motor parameter and clinical scales
Correlation analyses of motor parameters and clinical severity are shown in Table 4. For the total SARA score, positive correlations were observed for arrhythmicity in SIP (r = 0.587, P = 0.001, Fig. 2A), and trunk deflection range in the mediolateral direction during walking (r = 0.561, P = 0.002, Fig 2B). The SARA gait subscore correlates showed positive correlation in trunk deflection range in the mediolateral direction (ρ = 0.556, P = 0.002, Fig. 2C). For the SARA subscores for stance, sitting, heel-shin slide, there were no motor parameter found for correction, after Bonferroni correction.
The main findings of the current study are as follows: Step length during walking and anteroposterior truncal sway while standing are both associated with normal aging and body height. Gait speed was slower by normal aging. Analysis of the differences in gait between patients with CA and healthy controls identified slower gait speed, greater step width, and decreased step length when walking at comfortable speed in the CA group. Amplitude of the mediolateral truncal sway while walking was significantly increased in CA group when compared with HC group and was correlated with the SARA total score, and the only motor parameter which corelated with SARA gait subscore. When stepping in place, arrhythmicity of stepping and stance time increased in the CA group, in which the former correlated with the SARA total score. Abnormal increase in mediolateral trunk sway in closed-feet stance with eyes open was found in CA group.
Gait parameters
Individuals with balance deficits exhibit gait characteristics that bear similarities to cerebellar gait ataxia. These include decreased stride lengths, increased stride widths, decreased joint range of motion, abnormal free leg trajectories, greater variability in foot placement and decomposition, and reduced walking speed [29, 30]. In contrast, participants with impaired leg coordination do not show gait abnormalities [6], indicating that the characteristics of cerebellar gait ataxia are attributed to imbalance. This is consistent with our result that while two out of 12 motor parameters showed correlations to SARA total and SARA gait subscores, no correlation was observed with SARA heel-shin slide subscore. The decrease in dynamic balance and stability associated with CA can be assumed to increase the center of mass variability. To compensate for this, individuals may adopt a wide-based gait, thereby expanding the basal plane of support. Step width during walking was the only parameter observed in the present study that correlated with both the SARA gait and stance subscores.
Another compensatory strategy to avoid balance loss is to reduce the duration of the single-support phase, which is the most unstable body configuration [6]. In line with this, stance time in the SIP task was prolonged in the present study. In addition to step width, features such as step length and gait speed may also reflect the underlying compensatory mechanisms for ataxic gait, as they correlated with each other (P < 0.05, data not shown).
Increased mediolateral amplitude of truncal sway while walking, in line with a previous study [31], may more directly reflect impairment in maintaining truncal balance. Furthermore, the amplitude of the trunk in the lateral direction most prominently correlated with the SARA gait subscore and may serve as a biomarker for CA gait. Examination of truncal sway during walking is difficult through visual assessment; therefore, automated motion analysis is an efficient tool to assess individuals with CA.
Postural balance
Lateral postural control is important for stabilizing the standing position, as factors in the lateral direction have a more significant impact on instability than those in the anteroposterior direction [1, 32, 33]. Although the trunk sway while standing in this study did not correlate with any SARA scores, it does not refute the clinical usefulness of truncal sway. In contrast, the postural control assessments conducted in the closed feet stance using automated computerized analysis may offer subtle results that may be undetectable through visual inspection. Such measurements may also have a role in screening for pathology.
A previous study using accelerometers to assess standing in patients with spinocerebellar degeneration reported an increase in trunk sway in both the anteroposterior and lateral directions [34]. Another report showed that the amplitude of truncal sway increases in the mediolateral direction during standing, which is in accordance with the present study findings [29]. The directionality of instability is dependent on stance width, with decreased lateral sway upon increasing stance width in SCA6 [35] and alcoholism [36]. Therefore, the stance width should ideally be standardized and reported with results, such as natural stance, shoulder width stance or stance standardized with foot template. Choosing closed stance as a standard condition as in our study may result in a dropout at the later stage of the disease, which may be problematic in a longitudinal observation study. Nonetheless, it may be efficacious in a short-term observation. In addition, the positive correlation between anteroposterior truncal sway and aging may have contributed to the results that did not show any differences in this parameter. Importantly, the main feature of increased mediolateral sway while standing was not dependent on age or body stature in our HC sample.
Variability of lower limb movements by stepping in place
Increased spatiotemporal gait variability has been reported as the most consistent feature of gait patterns in CA [37–40]. Gait analysis of spinocerebellar degeneration using a comprehensive set of gait measurements with body-worn sensors showed that the standard deviations in the toe-out angle, toe-off angle, double-support time, and foot elevation at mid-swing were higher than those in healthy individuals [16]. The standard deviations in double-support time and elevation of the feet at mid-swing positively correlated with the SARA score severity [16]. The temporal and spatial increases in gait variability correlated with disease duration in SCA1 and SCA6, but not in SCA2 and SCA3, which may be owed to neurological signs other than ataxia affecting gait at different disease stages [16].
Other researchers claim that temporal, but not spatial, variability is a consistent abnormal feature of CA [30]. It has been demonstrated that variability between strides is a risk factor for falling and is greater in the older individuals than in younger individuals [7, 41, 42]. Gait measurement using Vicon cameras to examine the intra-limb coordination patterns of the knee, hip, and ankle joints demonstrated increased temporal variability of intra-limb coordination in patients with CA compared with that in patients with Parkinson's disease and vestibular disorders, and this result was independent of decreased walking speed [39]. The temporal variability of intra-limb coordination highly correlated with the clinical subscale of motor disability, indicating impairment in the control of limb movements and intra-limb coordination. However, this spatial variability also increased in patients with Parkinson's disease; therefore, it may not be specific to patients with CA [39]. Gait variability provides indirect information regarding the neural control of movement [43]. This variability has been implicated in the ability to maintain dynamic balance through a regular gait pattern and may serve as an early indicator of falls.
The total SARA score correlated both with truncal lateral sway while walking and arrhythmicity in SIP, which represents the variation in the step time. Although acquired from SIP as a non-locomotor stepping task, arrhythmicity was correlated with the gait parameters from SCSW such as, mediolateral deflection range of the trunk during walking, gait speed, and step length while walking (data not shown). This suggests that arrhythmicity during stepping as a core feature of gait abnormality in patients with CA can be effectively captured in stepping in place paradigm.
Bruijn et al. stated that at least 150 strides are required to achieve sufficient accuracy in the assessment of motor parameters during gait [44]. This is not feasible in clinical practice, and many authors have used 20–35 strides, which are suitable for adequately estimating gait stability in healthy individuals and patients [45, 46]. Establishing a measurement environment that offers sufficient space for multiple motion capture devices while minimizing interruptions from individuals or obstacles is a challenging task. Therefore, the advantages of performing SIP are that it is practical (as it does not require a large area), allows continuous measurements, and evaluates the results in a simple manner rather than using complex algorithms. Further validation against gait variability obtained from walking tests is currently under investigation.
Study limitations
The current study has several limitations. First, because of the nature of the measurement, the degree of motor disability that could be evaluated was limited to ambulatory individuals who did not require assistance. Therefore, the disease severity was mild, and the range of severity was inevitably small, with respect to the whole disease process. In addition, cross-sectional observations may not be suitable for the mild CA group because of the high variability, probably owing to compensatory mechanisms. Nevertheless, motor assessment could quantitatively detect abnormalities and their severity in early phase of the disease, which is a likely target for the future intervention study. A larger study population is required to clarify the relationship between brain function and ataxia severity. Second, patients with hereditary spinocerebellar degeneration and MSA-C were included in the CA group. There were no group differences (P > 0.20), except for mediolateral trunk sway in the closed feet stance, which was higher in the MSA-C group (P = 0.006). We cannot exclude the possibility that patients with MSA may have extrapyramidal and autonomic systems that may influence their movements; hence, further investigation of MSA and hereditary spinocerebellar degeneration is warranted to investigate effects of non-cerebellar involvement. Another limitation is the lack of patient-oriented reports. Further evaluation, such as patient-report of awareness of problem and activities of daily living function, may address further our understanding of CA.
The gait and postural control analyses demonstrated quantitative and concise motor deficits in patients with CA. The advantages of using this three-dimensional automatic motion analysis are that it is simple to conduct and can be objectively and quantitatively analyzed with high reproducibility and reliability. The mediolateral trunk deflection range during walking and arrhythmicity while stepping in space are candidate digital motor biomarkers for the diagnosis and symptomatic assessment of patients with CA.
Acknowledgements
We thank all the participants for their contributions to this study. We are also grateful to the staff members of the Division of Rehabilitation Medicine, Chiba University Hospital, for their technical support.
Ethical approval
This study was approved by the Ethics Committee of the Graduate School of Medicine, Chiba University (approval no. 3174). Written informed consent for participation and publication was obtained.
Competing interests
KO and AUB are shareholders of Motognosis GmbH and are named inventors of patent applications describing perceptive visual computing for tracking motor dysfunction. HMR was employed at Motognosis GmbH. All other authors have no conflicts of interest to declare relevant to the content of this article.
Authors’ contributions
MS: data acquisition, writing manuscript; SH; project administration, participant administration, data curation, writing manuscript; KO: Data acquisition and analysis, software, editing manuscript; TSH: study design, editing manuscript; MI, MT; data acquisition and curation; RK, AS: participant administration, editing manuscript; MM: project management, editing manuscript HMR: Data curation, software, reviewing manuscript; AUB, AM, FP, SK: supervision and review manuscript
Funding
No funding.
Availablity of data and materials
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
- Manto M and Marmolino D. Cerebellar ataxias. Curr Opin Neurol 2009: 22:419–29.
- Maetzler W, Domingos J, Srulijes K, Ferreira JJ and Bloem BR. Quantitative wearable sensors for objective assessment of Parkinson's disease. Mov Disord 2013: 28:1628–37.
- Galna B, Barry G, Jackson D, Mhiripiri D, Olivier P and Rochester L. Accuracy of the Microsoft Kinect sensor for measuring movement in people with Parkinson's disease. Gait Posture 2014: 39:1062–8.
- Behrens JR, Mertens S, Kruger T, Grobelny A, Otte K, Mansow-Model S, Gusho E, Paul F, Brandt AU and Schmitz-Hubsch T. Validity of visual perceptive computing for static posturography in patients with multiple sclerosis. Mult Scler 2016: 22:1596 – 606.
- Chini G, Ranavolo A, Draicchio F, Casali C, Conte C, Martino G, Leonardi L, Padua L, Coppola G, Pierelli F and Serrao M. Local Stability of the Trunk in Patients with Degenerative Cerebellar Ataxia During Walking. Cerebellum 2017: 16:26–33.
- Serrao M, Pierelli F, Ranavolo A, Draicchio F, Conte C, Don R, Di Fabio R, LeRose M, Padua L, Sandrini G and Casali C. Gait pattern in inherited cerebellar ataxias. Cerebellum 2012: 11:194–211.
- Ilg W and Timmann D. Gait ataxia–specific cerebellar influences and their rehabilitation. Mov Disord 2013: 28:1566–75.
- Hadjivassiliou M, Martindale J, Shanmugarajah P, Grunewald RA, Sarrigiannis PG, Beauchamp N, Garrard K, Warburton R, Sanders DS, Friend D, Duty S, Taylor J and Hoggard N. Causes of progressive cerebellar ataxia: prospective evaluation of 1500 patients. J Neurol Neurosurg Psychiatry 2017: 88:301–9.
- Tsuji S, Onodera O, Goto J, Nishizawa M and Study Group on Ataxic D. Sporadic ataxias in Japan–a population-based epidemiological study. Cerebellum 2008: 7:189–97.
- Sakai H, Yoshida K, Shimizu Y, Morita H, Ikeda S and Matsumoto N. Analysis of an insertion mutation in a cohort of 94 patients with spinocerebellar ataxia type 31 from Nagano, Japan. Neurogenetics 2010: 11:409–15.
- Buckley E, Mazza C and McNeill A. A systematic review of the gait characteristics associated with Cerebellar Ataxia. Gait Posture 2018: 60:154–63.
- Cabaraux P, Agrawal SK, Cai H, Calabro RS, Casali C, Damm L, Doss S, Habas C, Horn AKE, Ilg W, Louis ED, Mitoma H, Monaco V, Petracca M, Ranavolo A, Rao AK, Ruggieri S, Schirinzi T, Serrao M, Summa S, Strupp M, Surgent O, Synofzik M, Tao S, Terasi H, Torres-Russotto D, Travers B, Roper JA and Manto M. Consensus Paper: Ataxic Gait. Cerebellum 2023: 22:394–430.
- Schmitz-Hubsch T, du Montcel ST, Baliko L, Berciano J, Boesch S, Depondt C, Giunti P, Globas C, Infante J, Kang JS, Kremer B, Mariotti C, Melegh B, Pandolfo M, Rakowicz M, Ribai P, Rola R, Schols L, Szymanski S, van de Warrenburg BP, Durr A, Klockgether T and Fancellu R. Scale for the assessment and rating of ataxia: development of a new clinical scale. Neurology 2006: 66:1717–20.
- Yabe I, Matsushima M, Soma H, Basri R and Sasaki H. Usefulness of the Scale for Assessment and Rating of Ataxia (SARA). J Neurol Sci 2008: 266:164–6.
- Ilg W, Seemann J, Giese M, Traschutz A, Schols L, Timmann D and Synofzik M. Real-life gait assessment in degenerative cerebellar ataxia Toward ecologically valid biomarkers. Neurology 2020: 95:E1199-E210.
- Shah VV, Rodriguez-Labrada R, Horak FB, McNames J, Casey H, Hansson Floyd K, El-Gohary M, Schmahmann JD, Rosenthal LS, Perlman S, Velazquez-Perez L and Gomez CM. Gait Variability in Spinocerebellar Ataxia Assessed Using Wearable Inertial Sensors. Mov Disord 2021: 36:2922–31.
- Otte K, Kayser B, Mansow-Model S, Verrel J, Paul F, Brandt AU and Schmitz-Hubsch T. Accuracy and Reliability of the Kinect Version 2 for Clinical Measurement of Motor Function. PLoS One 2016: 11:e0166532.
- Muller B, Ilg W, Giese MA and Ludolph N. Validation of enhanced kinect sensor based motion capturing for gait assessment. PLoS One 2017: 12:e0175813.
- Summa S, Tartarisco G, Favetta M, Buzachis A, Romano A, Bernava GM, Vasco G, Pioggia G, Petrarca M, Castelli E, Bertini E and Schirinzi T. Spatio-temporal parameters of ataxia gait dataset obtained with the Kinect. Data Brief 2020: 32:106307.
- Otte K, Ellermeyer T, Suzuki M, Rohling HM, Kuroiwa R, Cooper G, Mansow-Model S, Mori M, Zimmermann H, Brandt AU, Paul F, Hirano S, Kuwabara S and Schmitz-Hubsch T. Cultural bias in motor function patterns: Potential relevance for predictive, preventive, and personalized medicine. EPMA J 2021: 12:91–101.
- Rohling HM, Otte K, Rekers S, Finke C, Rust R, Dorsch EM, Behnia B, Paul F and Schmitz-Hubsch T. RGB-Depth Camera-Based Assessment of Motor Capacity: Normative Data for Six Standardized Motor Tasks. Int J Environ Res Public Health 2022: 19.
- Behrens J, Pfuller C, Mansow-Model S, Otte K, Paul F and Brandt AU. Using perceptive computing in multiple sclerosis - the Short Maximum Speed Walk test. J Neuroeng Rehabil 2014: 11:89.
- Grobelny A, Behrens JR, Mertens S, Otte K, Mansow-Model S, Kruger T, Gusho E, Bellmann-Strobl J, Paul F, Brandt AU and Schmitz-Hubsch T. Maximum walking speed in multiple sclerosis assessed with visual perceptive computing. PLoS One 2017: 12:e0189281.
- Otte K, Ellermeyer T, Vater TS, Voigt M, Kroneberg D, Rasche L, Kruger T, Rohling HM, Kayser B, Mansow-Model S, Klostermann F, Brandt AU, Paul F, Lipp A and Schmitz-Hubsch T. Instrumental Assessment of Stepping in Place Captures Clinically Relevant Motor Symptoms of Parkinson's Disease. Sensors (Basel) 2020: 20.
- Abele M, Burk K, Schols L, Schwartz S, Besenthal I, Dichgans J, Zuhlke C, Riess O and Klockgether T. The aetiology of sporadic adult-onset ataxia. Brain 2002: 125:961–8.
- Gilman S, Wenning GK, Low PA, Brooks DJ, Mathias CJ, Trojanowski JQ, Wood NW, Colosimo C, Durr A, Fowler CJ, Kaufmann H, Klockgether T, Lees A, Poewe W, Quinn N, Revesz T, Robertson D, Sandroni P, Seppi K and Vidailhet M. Second consensus statement on the diagnosis of multiple system atrophy. Neurology 2008: 71:670–6.
- Plotnik M, Giladi N, Balash Y, Peretz C and Hausdorff JM. Is freezing of gait in Parkinson's disease related to asymmetric motor function? Ann Neurol 2005: 57:656–63.
- Schmitz-Hubsch T, Fimmers R, Rakowicz M, Rola R, Zdzienicka E, Fancellu R, Mariotti C, Linnemann C, Schols L, Timmann D, Filla A, Salvatore E, Infante J, Giunti P, Labrum R, Kremer B, van de Warrenburg BP, Baliko L, Melegh B, Depondt C, Schulz J, du Montcel ST and Klockgether T. Responsiveness of different rating instruments in spinocerebellar ataxia patients. Neurology 2010: 74:678 – 84.
- Morton SM and Bastian AJ. Relative contributions of balance and voluntary leg-coordination deficits to cerebellar gait ataxia. J Neurophysiol 2003: 89:1844–56.
- Morton SM and Bastian AJ. Mechanisms of cerebellar gait ataxia. Cerebellum 2007: 6:79–86.
- Shirai S, Yabe I, Matsushima M, Ito YM, Yoneyama M and Sasaki H. Quantitative evaluation of gait ataxia by accelerometers. J Neurol Sci 2015: 358:253–8.
- Lanzetta D, Cattaneo D, Pellegatta D and Cardini R. Trunk control in unstable sitting posture during functional activities in healthy subjects and patients with multiple sclerosis. Arch Phys Med Rehabil 2004: 85:279–83.
- Kavanagh JJ, Morrison S and Barrett RS. Coordination of head and trunk accelerations during walking. Eur J Appl Physiol 2005: 94:468–75.
- Van de Warrenburg BP, Bakker M, Kremer BP, Bloem BR and Allum JH. Trunk sway in patients with spinocerebellar ataxia. Mov Disord 2005: 20:1006–13.
- Bunn LM, Marsden JF, Giunti P and Day BL. Stance instability in spinocerebellar ataxia type 6. Mov Disord 2013: 28:510–6.
- Sullivan EV, Rose J and Pfefferbaum A. Physiological and focal cerebellar substrates of abnormal postural sway and tremor in alcoholic women. Biol Psychiatry 2010: 67:44–51.
- Palliyath S, Hallett M, Thomas SL and Lebiedowska MK. Gait in patients with cerebellar ataxia. Mov Disord 1998: 13:958–64.
- Ebersbach G, Sojer M, Valldeoriola F, Wissel J, Muller J, Tolosa E and Poewe W. Comparative analysis of gait in Parkinson's disease, cerebellar ataxia and subcortical arteriosclerotic encephalopathy. Brain 1999: 122 (Pt 7):1349–55.
- Ilg W, Golla H, Thier P and Giese MA. Specific influences of cerebellar dysfunctions on gait. Brain 2007: 130:786–98.
- Ilg W, Giese MA, Gizewski ER, Schoch B and Timmann D. The influence of focal cerebellar lesions on the control and adaptation of gait. Brain 2008: 131:2913–27.
- Barak Y, Wagenaar RC and Holt KG. Gait characteristics of elderly people with a history of falls: a dynamic approach. Phys Ther 2006: 86:1501–10.
- Hollman JH, Kovash FM, Kubik JJ and Linbo RA. Age-related differences in spatiotemporal markers of gait stability during dual task walking. Gait Posture 2007: 26:113–9.
- Hausdorff JM. Stride variability: beyond length and frequency. Gait Posture 2004: 20:304; author reply 5.
- Bruijn SM, Meijer OG, Beek PJ and van Dieen JH. Assessing the stability of human locomotion: a review of current measures. J R Soc Interface 2013: 10:20120999.
- van Schooten KS, Sloot LH, Bruijn SM, Kingma H, Meijer OG, Pijnappels M and van Dieen JH. Sensitivity of trunk variability and stability measures to balance impairments induced by galvanic vestibular stimulation during gait. Gait Posture 2011: 33:656–60.
- van Schooten KS, Rispens SM, Elders PJ, van Dieen JH and Pijnappels M. Toward ambulatory balance assessment: estimating variability and stability from short bouts of gait. Gait Posture 2014: 39:695–9.
Table 1 Descriptions of the three motor tasks and derived spatiotemporal parameters
|
Motor task |
Instruction |
Movement signals |
Derived clinical parameters |
|
Short comfortable speed walk (SCSW) |
After audio signal, walk directly toward the sensor at a comfortable speed |
Spatial body movement (spine base movement) and ankle movements in walk direction |
Average gait speed (m/s), cadence (steps/min), average step length (cm), average anteroposterior and mediolateral trunk deflection range (°) |
|
Stance with open eyes (POCO) |
Stand with closed feet and open eyes for 20 s |
Body sway (movement of spine base relative to closed feet position) |
Average anteroposterior and mediolateral trunk deflection range (°) |
|
Stepping in place (SIP) |
Walk on the spot at a comfortable pace for 40 s |
Antero-postero-vertical displacement of the knees |
Amplitude asymmetry (%), arrhythmicity (%), mean step time (s), mean stance time (s) |
Table 2 Descriptive demography of groups
|
Healthy control Mean (SD) |
Cerebellar ataxia a Mean (SD) |
P-value b |
|
|
|
|||
|
Sex (male:female) |
18:15 |
20:8 |
0.20 |
|
Age (years) |
58.6 (13.9) |
59.6 (10.8) |
0.88 |
|
Body height (cm) |
163.1 (9.1) |
165.6 (6.7) |
0.23 |
|
Body weight (kg) |
62.9 (9.5) |
62.4 (9.9) |
0.90 |
|
BMI (kg/m2) |
23.6 (2.9) |
22.7 (3.2) |
0.29 |
|
Disease duration (years) |
4.88 (4.48) |
||
|
SARA total score |
|
12.07 (2.94) |
|
|
item 1, gait |
|
2.79 (0.77) |
|
|
item 2, stance |
|
2.04 (0.63) |
|
|
item 3, sitting |
|
0.19 (0.47) |
|
|
item 8, heel-shin slide |
|
1.86 (0.71) |
|
BMI: body mass index; SARA: scale for the assessment and rating of ataxia; SD: standard deviation
a Including eight with spinocerebellar ataxia (SCA) type 6, three with SCA31, two with dominant-inherited cerebellar cortical atrophy, one with cortical cerebellar atrophy, and 14 with multiple system atrophy with predominant cerebellar ataxia.
b Chi-square test for sex differences and unpaired T test for other parameters.
Table 3 Descriptive statistics of 12 spatiotemporal motor parameters derived from three different motor tasks, and comparisons of patients with cerebellar ataxia and healthy controls.
|
Motor parameters |
Healthy controls |
Cerebellar ataxia |
P-value |
||
|
Mean (SD) |
CV |
Mean (SD) |
CV |
|
|
|
Short comfortable speed walk |
|
|
|
|
|
|
AP deflection range (°) |
5.15 (2.84) |
0.30 |
5.87 (1.78) |
0.54 |
0.19 |
|
ML deflection range (°) |
2.67 (0.86) |
0.29 |
3.85 (1.11) |
0.32 |
< 0.001 |
|
Step length (cm) |
63.6 (9.58) |
0.24 |
45.1 (11.0) |
0.15 |
< 0.001 |
|
Step width (cm) |
11.2 (2.33) |
0.26 |
17.1 (4.44) |
0.23 |
< 0.001 |
|
Gait speed (m/s) |
1.12 (0.19) |
0.26 |
0.80 (0.20) |
0.16 |
< 0.001 |
|
Cadence (steps/min) |
118 (8.88) |
0.10 |
114 (11.5) |
0.07 |
0.16 |
|
|
|
|
|||
|
Stance with closed feet |
|
|
|
|
|
|
AP deflection range (°) |
1.35 (0.73) |
0.46 |
1.28 (0.60) |
0.54 |
0.83 |
|
ML deflection range (°) |
1.01 (0.42) |
0.52 |
2.00 (1.04) |
0.41 |
< 0.001 |
|
|
|
|
|||
|
Stepping in place |
|
|
|
|
|
|
Asymmetry (%) |
0.04 (0.03) |
1.00 |
0.09 (0.09) |
0.70 |
0.012 a |
|
Arrhythmicity (%) |
7.26 (4.57) |
0.38 |
13.9 (5.28) |
0.61 |
< 0.001 |
|
Average step time (s) |
0.79 (0.06) |
0.15 |
0.77 (0.12) |
0.10 |
0.21 |
|
Average stance time (s) |
0.34 (0.07) |
0.52 |
0.55 (0.28) |
0.29 |
< 0.001 a |
SD: standard deviation; CV: coefficient of variation; AP: anteroposterior; ML: mediolateral
Comparison with healthy control group using unpaired t-test, except a: Mann–Whitney U test.
Table 4 Correlation analyses between motor parameters and ataxia clinical scale in 28 patients with cerebellar ataxia
|
Motor parameters |
Scale for the Assessment and Rating of Ataxia (SARA) |
||||
|
Total score |
Gait |
Stance |
Sitting |
Heel-shin |
|
|
Short comfortable speed walk |
|
|
|
|
|
|
AP deflection range (°) |
0.198 a |
0.024 |
-0.021 |
-0.140 |
0.084 |
|
ML deflection range (°) |
0.561* a |
0.556* |
0.132 |
0.062 |
0.275 |
|
Step length (cm) |
-0.453 a |
-0.434 |
-0.335 |
-0.481 |
-0.306 |
|
Step width (cm) |
0.405 a |
0.473 |
0.379 |
0.095 |
0.165 |
|
Gait speed (m/s) |
-0.332 a |
-0.444 |
-0.224 |
-0.448 |
-0.185 |
|
Cadence (steps/min) |
0.236 a |
-0.072 |
0.020 |
0.032 |
0.303 |
|
Stance with closed feet |
|
|
|
|
|
|
AP deflection range (°) |
0.373 a |
0.275 |
0.153 |
0.212 |
0.180 |
|
ML deflection range (°) |
-0.002 a |
0.054 |
0.200 |
0.265 |
-0.074 |
|
Stepping in place |
|
|
|
|
|
|
Asymmetry (%) |
0.087 |
0.309 |
0.043 |
0.175 |
0.063 |
|
Arrhythmicity (%) |
0.587*a |
0.426 |
0.394 |
0.226 |
0.252 |
|
Average step time (s) |
-0.241 a |
-0.230 |
-0.059 |
0.179 |
-0.221 |
|
Average stance time (s) |
0.085 |
0.150 |
0.059 |
0.159 |
-0.120 |
Shown in correlation coefficients (ρ) except a: (r).
AP: anteroposterior, ML: mediolateral
Assessments used were, a: Pearson’s correlation, otherwise Spearman’s rank correlation, if not mentioned.
Significance levels are indicated by asterisks (*P < 0.05 with Bonferroni correction).
Competing interest reported. KO and AUB are shareholders of Motognosis GmbH and are named inventors of patent applications describing perceptive visual computing for tracking motor dysfunction. HMR was employed at Motognosis GmbH. All other authors have no conflicts of interest to declare relevant to the content of this article.