Falls are an eminent risk for older adults and especially patients with neurodegenerative disorders, such as Parkinson's disease (PD). Recent advancements in wearable sensor technology and machine learning may provide a possibility for an individualized prediction of fall risk based on gait recordings from standardized gait tests or from unconstrained real-world scenarios. However, the most effective aggregation of continuous real-world data as well as the potential of unsupervised gait tests recorded over multiple days for fall risk prediction still need to be investigated. Therefore, we present a data set containing real-world gait and unsupervised 4x10-Meter-Walking-Tests of 40 PD patients, continuously recorded with foot-worn inertial sensors over a period of two weeks. In this prospective study, falls were self-reported during a three-month follow-up phase, serving as ground truth for fall risk prediction. The purpose of this study was to compare different data aggregation approaches and machine learning models for the prospective prediction of fall risk using gait parameters derived either from continuous real-world recordings or from unsupervised gait tests. The highest balanced accuracy of 74.0% (sensitivity: 60.0%, specificity: 88.0%) was achieved with a Random Forest Classifier applied to the real-world gait data when aggregating all walking bouts and days of each participant. Our findings suggest that fall risk can be predicted best by merging the entire two-week real-world gait data of a patient, outperforming the prediction using unsupervised gait tests (68.0% balanced accuracy) and contribute to an improved understanding of fall risk prediction.

Digital mobility assessment using wearable sensor systems has the potential to capture walking performance in a patient’s natural environment. It enables monitoring of health status and disease progression and evaluation of interventions in real-world situations. In contrast to laboratory settings, real-world walking occurs in non-conventional environments and under unconstrained and uncontrolled conditions. Despite the general understanding, there is a lack of agreed definitions about what constitutes real-world walking, impeding the comparison and interpretation of the acquired data across systems and studies. The goal of this study was to obtain expert-based consensus on specific aspects of real-world walking and to provide respective definitions in a common terminological framework. An adapted Delphi method was used to obtain agreed definitions related to real-world walking. In an online survey, 162 participants from a panel of academic, clinical and industrial experts with experience in the field of gait analysis were asked for agreement on previously specified definitions. Descriptive statistics was used to evaluate whether consent (> 75% agreement as defined a priori) was reached. Of 162 experts invited to participate, 51 completed all rounds (31.5% response rate). We obtained consensus on all definitions (“Walking” > 90%, “Purposeful” > 75%, “Real-world” > 90%, “Walking bout” > 80%, “Walking speed” > 75%, “Turning” > 90% agreement) after two rounds. The identification of a consented set of real-world walking definitions has important implications for the development of assessment and analysis protocols, as well as for the reporting and comparison of digital mobility outcomes across studies and systems. The definitions will serve as a common framework for implementing digital and mobile technologies for gait assessment and are an important link for the transition from supervised to unsupervised gait assessment.

Running has a positive impact on human health and is an accessible sport for most people. There is high demand for tracking running performance and progress for amateurs and professionals alike. The parameters velocity and distance are thereby of main interest. In this work, we evaluate the accuracy of four algorithms, which calculate the stride velocity and stride length during running using data of an inertial measurement unit (IMU) placed in the midsole of a running shoe. The four algorithms are based on stride time, foot acceleration, foot trajectory estimation, and deep learning, respectively. They are compared using two studies: a laboratory-based study comprising 2377 strides from 27 subjects with 3D motion tracking as a reference and a field study comprising 12 subjects performing a 3.2-km run in a real-world setup. The results show that the foot trajectory estimation algorithm performs best, achieving a mean error of 0.032 ± 0.274 m/s for the velocity estimation and 0.022 ± 0.157 m for the stride length. An interesting alternative for systems with a low energy budget is the acceleration-based approach. Our results support the implementation decision for running velocity and distance tracking using IMUs embedded in the sole of a running shoe.

Gait tests as part of home monitoring study protocols for patients with movement disorders may provide valuable standardized anchor-points for real-world gait analysis using inertial measurement units (IMUs). However, analyzing unsupervised gait tests relies on reliable test annotations by the patients requiring a potentially error-prone interaction with the recording system. To overcome this limitation, this work presents a novel algorithmic pipeline for the automated detection of unsupervised standardized gait tests from continuous real-world IMU data. In a study with twelve Parkinson’s disease patients, we recorded real-world gait data over two weeks using foot-worn IMUs. During continuous daily recordings, the participants performed series of three consecutive ${4}\times {10}$ -Meters-Walking-Tests ( ${4}\times {10}$ MWTs) at different walking speeds, besides their usual daily-living activities. The algorithm first detected these gait test series using a gait sequence detection algorithm, a peak enhancement pipeline, and subsequence Dynamic Time Warping and then decomposed them into single ${4}\times {10}$ MWTs based on the walking speed. In the evaluation with 419 available gait test series, the detection reached an F1-score of 88.9% and the decomposition an F1-score of 94.0%. A concurrent validity evaluation revealed very good agreement between spatio-temporal gait parameters derived from manually labelled and automatically detected ${4}\times {10}$ MWTs. Our algorithm allows to remove the burden of system interaction from the patients and reduces the time for manual data annotation for researchers. The study contributes to an improved automated processing of real-world IMU gait data and enables a simple integration of standardized tests into continuous long-term recordings. This will help to bridge the gap between supervised and unsupervised gait assessment.

Background: To objectively assess a patient’s gait, a robust identification of stride borders is one of the first steps in inertial sensor-based mobile gait analysis pipelines. While many different methods for stride segmentation have been presented in the literature, an out-of-lab evaluation of respective algorithms on free-living gait is still missing. Method: To address this issue, we present a comprehensive free-living evaluation dataset, including 146.574 semi-automatic labeled strides of 28 Parkinson’s Disease patients. This dataset was used to evaluate the segmentation performance of a new Hidden Markov Model (HMM) based stride segmentation approach compared to an available dynamic time warping (DTW) based method. Results: The proposed HMM achieved a mean F1-score of 92.1% and outperformed the DTW approach significantly. Further analysis revealed a dependency of segmentation performance to the number of strides within respective walking bouts. Shorter bouts ( $$< 30$$< 30 strides) resulted in worse performance, which could be related to more heterogeneous gait and an increased diversity of different stride types in short free-living walking bouts. In contrast, the HMM reached F1-scores of more than 96.2% for longer bouts ( $$> 50$$> 50 strides). Furthermore, we showed that an HMM, which was trained on at-lab data only, could be transferred to a free-living context with a negligible decrease in performance. Conclusion: The generalizability of the proposed HMM is a promising feature, as fully labeled free-living training data might not be available for many applications. To the best of our knowledge, this is the first evaluation of stride segmentation performance on a large scale free-living dataset. Our proposed HMM-based approach was able to address the increased complexity of free-living gait data, and thus will help to enable a robust assessment of stride parameters in future free-living gait analysis applications.

Mobile gait analysis using wearable inertial measurement units (IMUs) provides valuable insights for the assessment of movement impairments in different neurological and musculoskeletal diseases, for example Parkinson's disease (PD). The increase in data volume due to arising long-term monitoring requires valid, robust and efficient analysis pipelines. In many studies an upstream detection of gait is therefore applied. However, current methods do not provide a robust way to successfully reject non-gait signals. Therefore, we developed a novel algorithm for the detection of gait from continuous inertial data of sensors worn at the feet. The algorithm is focused not only on a high sensitivity but also a high specificity for gait. Sliding windows of IMU signals recorded from the feet of PD patients were processed in the frequency domain. Gait was detected if the frequency spectrum contained specific patterns of harmonic frequencies. The approach was trained and evaluated on 150 clinical measurements containing standardized gait and cyclic movement tests. The detection reached as sensitivity of 0.98 and a specificity of 0.96 for the best sensor configuration (angular rate around the medio-lateral axis). On an independent validation data set including 203 unsupervised, semi-standardized gait tests, the algorithm achieved a sensitivity of 0.97. Our algorithm for the detection of gait from continuous IMU signals works reliably and showed promising results for the application in the context of free-living and non-standardized monitoring scenarios.