07/2019 – 01/2020
Huntington´s disease (HD) is a rare autosomal-dominant, neurodegenerative disease with various symptoms including movement impairment, cognitive as well as psychiatric disturbances. The predominant movement impairment is chorea, which is defined as involuntary, unpredictable body movements. The onset of symptoms typically appears around the age of 40 years . In clinical routine, motor impairment is measured by the Unified Huntington´s Disease Rating Scale (UHDRS) Total Motor Score (TMS), but it is often missed in early stages and the outcome of this rating scale is limited to interrater variability.
To support clinical decision making and enable a more objective assessment of movement impairment in neurodegenerative diseases, wearable sensors are becoming of increased. These sensors usually contain an inertial measurement unit (IMU) and can be used in clinic as well as in home-monitoring scenarios where clinicians usually do not have any insight into patient´s disease progression.
While there has been lots of research in the past years in Parkinson´s disease regarding the use of wearable sensors, there have been only few studies in HD. However, those studies already showed promising results concerning the feasibility of using motion sensors to classify HD patients and to determine the disease state with regard to the UHDRS.
Andrzejewski et al.  as well as Dinesh et al.  successfully classified HD patients vs controls as well as different disease stages based on a gait analysis using data of a single accelerometer worn at the chest or additional data of wearable accelerometers at the lower limbs. Classification accuracy was even higher regarding the data captured at home in contrast to the data measured in the clinic. Dalton et al.  also analysed the gait of HD patients as well as balance using a single accelerometer sensor worn at the thorax to differentiate between remanifest and manifest HD patients.
In other studies, upper limb movement impairment was considered to classify HD patients vs controls and to determine the disease state. Bennasar et al.  achieved these aims using data recorded from wearable accelerometers at both wrists and the chest while subjects perform the money box test. In contrast, Reilmann et al.  do so by using sensors to measure position and orientation of an instrument hold stable by the subjects. Lipsmeier et al.  showed significant differences between HD patients and controls in motion data including active tests and passive monitoring recorded by a test suite including a smartphone and a smartwatch.
For this thesis, similar to the approaches described in the literature, multiple body worn IMU sensors will be used to automatically distinguishing between HD patients and healthy subjects. The used sensor system will consist of five synchronized IMU sensors, one positioned at each limb and one hip sensor. This setting will allow to assess movement impairment of almost the whole body which have not been done so far. Additionally, we will not only collect data from accelerometer based sensors as previous studies did, but also from gyroscopes. A study initiated by the Department of Molecular Neurology of the UK Erlangen in collaboration with the MaDLab of the FAU Erlangen will collect motion data of HD patients and healthy subjects while performing tasks of a predefined protocol. In contrast to previous studies that focused either on static or motion tasks, this protocol includes static and non-static tasks as well as dual tasks with a cognitive component.
Based on the acquired dataset different classifiers will be developed and evaluated according to their performance in distinguishing between HD patients of various disease stages and healthy controls. In a second step this thesis should additionally face the task to determine the state of disease based on IMU sensor recorded motion data. The implementation should then be evaluated regarding the UHDRS as a reference.
This master thesis is embedded in an existing project of the department of Molecular Neurology of the UK Erlangen and the MaDLab to investigate clinical sensitive biomarkers for neurodegenerative diseases based on technology.
- Nance, M. A. (1998) “Huntington Disease: Clinical, Genetic, and Social Aspects” Journal of Geriatric Psychiatry and Neurology, vol. 11, issue 2, pp. 61-70
- Andrzejewski, K. L., et al. (2016) “Wearable Sensors in Huntington Disease: A Pilot Study.” Journal of Huntington´s Disease, vol. 5, no. 2, pp. 199-206
- Dinesh, K., Xiong, M., Adams, J., Dorsey, R., Sharma, G. () “Signal Anasysis for Detecting Motor Symptoms in Parkinson´s and Huntington´s Disease using musliple body-affixed Sensors: A Pilot Study.”
- Dalton, A., Khalil, H., Busse, M., Rosser, A., van Deursen, R., ÓLaighin, G.,(2012) “Analysis of gait and balance through a single triaxial accelerometer in presymptomatic and symptomatic Huntington´s disease” Gait and Posture, July 2012
- Bennasar, E., Hicks, Y. A., Clinch, S. P., Jones, P., Holt, C., Rossner, A., Busse, M. (2018). “Automated Assessment of Movement Impairment in Huntington´s Disease.” IEEE Transactions on Neural Systems and Rehabilitation Engineering