Parkinson’s disease is a neurodegenerative disorder impacting patients’ movement, causing a variety of movement abnormalities. It has been the focus of research studies for early detection based on wearable technologies. The benefit of wearable technologies in the domain rises by continuous monitoring of this population’s movement patterns over time. The ubiquity of wrist-worn accelerometry and the fact that the wrist is the most common and acceptable body location to wear the accelerometer for continuous monitoring suggests that wrist-worn accelerometers are the best choice for early detection of the disease and also tracking the severity of it over time. In this study, we use a dataset consisting of one-week wrist-worn accelerometry data collected from individuals with Parkinson’s disease and healthy elderlies for early detection of the disease. Two feature engineering methods, including epoch-based statistical feature engineering and the document-of-words method, were used. Using various machine learning classifiers, the impact of different windowing strategies, using the document-of-words method versus the statistical method, and the amount of data in terms of number of days were investigated. Based on our results, PD was detected with the highest average accuracy value (85% ± 15%) across 100 runs of SVM classifier using a set of features containing features from every and all windowing strategies. We also found that the document-of-words method significantly improves the classification performance compared to the statistical feature engineering model. Although the best performance of the classification task between PD and healthy elderlies was obtained using seven days of data collection, the results indicated that with three days of data collection, we can reach a classification performance that is not significantly different from a model built using seven days of data collection.

The fractal analysis of stride-to-stride fluctuations in walking has become an integral part of human gait research. Fractal analysis of stride time intervals can provide insights into locomotor function and dysfunction, but its application requires a large number of strides, which can be difficult to collect from people with movement disorders such as Parkinson’s disease. It has recently been suggested that “stitching” together short gait trials to create a longer time series could be a solution. The objective of this study was to determine if scaling exponents from “stitched” stride time series were similar to those from continuous, longer stride time series. Fifteen young adults, fourteen older adults, and thirteen people with Parkinson’s disease walked around an indoor track in three blocks: one time 15 min, five times 3 min, and thirty times 30 s. Stride time intervals were determined from gait events recorded with instrumented insoles, and the detrended fluctuation analysis was applied to each stride time series of 512 strides. There was no statistically significant difference between scaling exponents in the three blocks, but intra-class correlation revealed very low between-blocks reliability of scaling exponents. This result challenges the premise that the stitching procedure could provide reliable information about gait dynamics, as it suggests that fractal analysis of stitched time series does not capture the same dynamics as gait recorded continuously. The stitching procedure cannot be considered as a valid alternative to the collection of continuous, long trials. Further studies are recommended to determine if the application of fractal analysis is limited by its own methodological considerations (i.e., long time series), or if other solutions exists to obtain reliable scaling exponents in populations with movement disorders.