Using wearable sensors to evaluate turning in Parkinson’s disease: understanding freezing of gait and falls

Recent advances in the acquisition of motion signals have led to the development of portable, low-cost and reliable instruments. From them it is possible to extract objective and quantitative measures that can complement the traditional qualitative evaluations. As the population is getting older, new measurement systems are needed to assess what are the risks this population can encounter and which diseases it may develop. This thesis has the aim to evaluate turns by processing the signals obtained from a wearable sensor placed at the lumbar level of the patients. To achieve this goal, it has been decided to collect the quantitative signals obtained from three different studies and process them with the same turning algorithm to obtain a large dataset of digital mobility outcomes (DMOs). These are put all together to obtain a new complete database that will be used for classification and clinical validation analysis in an older population and in patients affected by Parkinson’s disease (PD). This work will focus on two of the most invalidating aspects that characterize motor symptoms in PD: falls and freezing of gait (FOG). Falling is considered one of the major causes of injury in the elderly population and can manifest in the PD population as consequence of gait impairment. FOG consists in the inability to move and can be experienced, together with other symptoms, by people who suffer from PD. This thesis has the aim to classify between fallers and non-fallers and between freezers and non-freezers by performing ROC analysis and by applying machine learning algorithms on qualitative and quantitative measurements of turning. The last objective is to clinically validate the DMOs by finding a relationship between them and demographic characteristics and clinical outcomes. The results highlight the importance of collecting quantitative measures that can allow accurate classification of the patient’s status by bringing multiple advantages for himself and clinicians.