Ambulatory assessment of motor performance after stroke
Stroke is the leading cause of serious and long-term disabilities in developed countries. It is a disruption of blood supply to brain tissue which results in irreversible cell death. Depending on size and position of the affected region, this disruption may cause many variations of disabilities. Of those who survive a stroke, common disabilities are upper and lower extremity motor deficits, cognitive dysfunction, incontinence and dysphagia and dysphasia. This dissertation focuses on this first category, i.e. deficits in motor performance caused by a stroke.
Despite intensive training programs, stroke survivors regularly show deterioration of motor function during or after their rehabilitation period. This deterioration may results in an expensive rehospitalization or institutionalization. In order to explain any functional progress or decline of motor function, information on body movements performed during daily life is necessary. Standardized lab-based systems for the evaluation of body movements are not suitable for measurements in a daily life setting. Systems, which are less obtrusive, wireless and wearable, may be more useful in this setting. However, today's wearable movement monitoring systems merely evaluate the functional performance during daily life activities quantitatively on an activity level (i.e., frequency and type of performed activities), and not qualitatively on a level of body function (i.e. how stroke survivors performed their movements). Although this information on the level of body function may be necessary to describe any progress or deterioration of motor performance during daily life activities.
Between November 2011 and January 2015, the international research team of the project: ‘INTERACTION’ developed a wearable sensing system for the ambulatory monitoring and assessment of upper and lower extremity motor performance. This system was developed based on inputs from questionnaires, interviews and consensus meetings with stroke survivors, health care professionals, engineers and researchers. The final version of the system consists of multiple sensors integrated in clothing and shoes.
Challenging in monitoring body movements in a daily life setting over a longer period of time, such as done by the INTERACTION system, is the complexity and the amount of data it generates. Health care professionals generally only have a couple of minutes per patient to evaluate all measurement data and to turn it into clinically relevant information. For that reason, metrics are needed that outline all data into concise functional and clinically relevant information, allowing the health care professional to interpret the data. The main goal of this dissertation therefore was to develop clinically relevant metrics for the assessment of stroke survivors' quality of upper and lower extremity motor performance. These metrics were estimated from on-body sensor data, measured using the INTERACTION sensing system, while stroke survivors performed activities of daily living. The methods presented in this dissertation were divided in three parts, which describe the evaluation of upper and lower extremity function, and the processing of the data into concise and functional measurement reports.
The methods presented in the first two parts of this dissertation, allow the assessment of motor performance during functional tasks in a daily life setting using a wearable sensor system such as the INTERACTION system. However, it remains difficult and time consuming for health care professionals to interpret this data, due to its large size. In the last part of this dissertation, data processing methods are proposed to present all measured data into concise functional reports. These reports present higher-level metrics within one figure, i.e., metrics that summarize multiple (repeated) body movements measured over a longer period of time. The automatically generated reports make it more feasible for the health care professional to evaluate large sets of movement data. In future, these reports may be used to objectively describe any progress or deterioration in motor performance.
The newly developed data analysis processes, described in this dissertation, enable the assessment of motor performance on a level of body function during functional tasks in a daily life setting. All presented methods were evaluated under controlled conditions and in simulated daily life settings. Further research should focus on the feasibility, usability and clinical implications of measuring daily life movements of stroke survivors in a daily life setting. The presented methods are not limited to the evaluation of motor performance in stroke survivors, but could also be used in other clinical cases to objectively evaluate motor performance and optimize individual rehabilitation processes.
