Veen, A.L. van der (Adriella) (2017) Evaluation of a phenotypic quantitative classification method for diagnosis of ataxia. thesis, Human Movement Sciences.
Full text available on request.Abstract
Ataxia is a symptom of movement disorders that affects balance and coordination of intentional movement. Ataxia can be diagnosed as a disorder by itself but the characteristic symptoms can also be present in other disorders. There are many variations in the clinical phenotype of ataxia and correct identification of these phenotypes can be difficult. Gait analysis can be sensitive for early identification of different phenotypes since it provides information on spatiotemporal characteristics of gait. In this study, we employed inertial sensors and a supervised classifier to obtain an automatic classification of ‘pure ataxia’ patients, ‘combined ataxia’ patients, and healthy controls. Data were recorded from sensors during gait in adults and children with ataxia and healthy controls. Based on signals from shank and thigh sensors, spatiotemporal features were determined and used as input for the classification of single strides as well as participants. The random forest method was used as classification method and cross-validation methods were used to estimate the performance of the classifier on a new dataset. Results revealed an overall accuracy of classifying strides of 87.1% and an accuracy of classifying participants of 54.7%. These results show that an automatic classifier based on spatiotemporal gait features is able to distinguish between pure ataxia, combined ataxia, and healthy controls at individual stride level. However, further research is necessary to obtain accurate classifications at patient level.
| Item Type: | Thesis (UNSPECIFIED) |
|---|---|
| Supervisor name: | Kamsma, Y.P.T. and Martinez-Manzanera, O.E. and Maurits, N.M. |
| Faculty: | Medical Sciences |
| Date Deposited: | 11 May 2022 09:38 |
| Last Modified: | 11 May 2022 09:39 |
| URI: | https://umcg.studenttheses.ub.rug.nl/id/eprint/3273 |
Actions (login required)
 |
View Item |