Activity Recognition for Persons With Stroke Using Mobile Phone Technology: Toward Improved Performance in a Home Setting

Pain Assessment in Osteoarthritis: Present Practices and Future Prospects Including the Use of Biomarkers and Wearable Technologies, and AI-Driven Personalized Medicine

Osteoarthritis (OA) is a highly prevalent chronic joint disorder affecting ~600 million individuals worldwide and is characterized by complex pain mechanisms that significantly impair patient quality of life. Challenges exist in accurately assessing and measuring pain in OA due to variations in pain perception among individuals and the heterogeneous nature of the disease. Conventional pain assessment methods, such as patient-reported outcome measures and clinical evaluations, often fail to fully capture the heterogeneity of pain experiences among individuals with OA. This review will summarize and evaluate current methods of pain assessment in OA and highlight future directions for standardized pain assessment. We discuss the role of animal models in enhancing our understanding of OA pain pathophysiology and highlight the necessity of translational research to advance pain assessment strategies. Key challenges explored include identifying phenotypes of pain susceptibility, integrating biomarkers into clinical practice, and adopting personalized pain management approaches through the incorporation of multi-modal data and multilevel analysis. We underscore the imperative for continued innovation in pain assessment and management to improve outcomes for patients with OA.

Determining the Prioritization of Behavior Change Techniques for Long-Term Stroke Rehabilitation: Delphi Survey Study

Background

Stroke results in both physical disability and psychological distress. The impact can be minimized through rehabilitation, but it is a long-term process, making it difficult for patients to adhere to treatment. Thus, a better understanding of long-term behavior change interventions for patients with stroke is needed as well as how such interventions can support not only rehabilitation of motoric functions but also mental well-being.

Objective

The aim of this study is to understand both the most important behavior change technique (BCT) clusters for long-term stroke rehabilitation in general as well as which are most relevant for each aspect of stroke rehabilitation: behavioral, cognitive, and emotional.

Methods

We applied the 16 BCT clusters. The study used a 2-round Delphi survey, as reliable consensus was obtained among a group of 12 international experts. Experts represented three main backgrounds involved in behavioral intervention in the health context: (1) specialists in behavioral science (n=4), (2) behavioral designers (n=4), and (3) expert health care professionals (n=4). Experts were brought together in this way for the first time. In the first round, web-based questionnaires were used to collect data from the experts. This was followed by a personalized second round. Consensus was determined by statistically aggregating the responses and evaluating IQR and percentage consensus. BCT clusters reaching consensus (IQR ≤1 and percentage ≥50%) were then ranked.

Results

In total, 12 of 16 BCT clusters reached consensus for general importance in stroke rehabilitation, with 11, 9, and 6 BCT clusters achieving consensus for, respectively, the behavioral, cognitive, and emotional aspects of rehabilitation. The overall most relevant BCT clusters were repetition and substitution, social support, feedback and monitoring, and self-belief, with similar outcomes for behavioral and cognitive rehabilitation. For emotional rehabilitation, social support and identity were emphasized. The least relevant BCT clusters were natural consequences, covert learning, and comparison of behavior.

Conclusions

This expert panel study using a 2-round Delphi survey ranked the importance of BCT clusters for long-term stroke rehabilitation. The process yielded a number of novel insights highlighting differences in importance between general rehabilitation and that specifically focused on the behavioral, cognitive, and emotional aspects of stroke recovery. This provides a first but important step toward unlocking the prioritization of BCT clusters for long-term intervention contexts such as stroke rehabilitation and enables effective intervention mapping addressing long-term behavior change and treatment adherence.

A systematic review of portable technologies for the early assessment of motor development in infants

Early screening and evaluation of infant motor development are crucial for detecting motor deficits and enabling timely interventions. Traditional clinical assessments are often subjective, without fully capturing infants' "real-world" behavior. This has sparked interest in portable, low-cost technologies to objectively and precisely measure infant motion at home, with a goal of enhancing ecological validity. In this systematic review, we explored the current landscape of portable, technology-based solutions to assess early motor development (within the first year), outlining the prevailing challenges and future directions. We reviewed 66 publications, which utilized video, sensors, or a combination of technologies. There were three key applications of these technologies: (1) automating clinical assessments, (2) illuminating new measures of motor development, and (3) predicting developmental outcomes. There was a promising trend toward earlier and more accurate detection using portable technologies. Additional research and demographic diversity are needed to develop fully automated, robust, and user-friendly tools. Registration & Protocol OSF Registries https://doi.org/10.17605/OSF.IO/R6JAE .

Sensor-Based Gait and Balance Assessment in Healthy Adults: Analysis of Short-Term Training and Sensor Placement Effects

While the analysis of gait and balance can be an important indicator of age- or disease-related changes, it remains unclear if repeated performance of gait and balance tests in healthy adults leads to habituation effects, if short-term gait and balance training can improve gait and balance performance, and whether the placement of wearable sensors influences the measurement accuracy. Healthy adults were assessed before and after performing weekly gait and balance tests over three weeks by using a force plate, motion capturing system and smartphone. The intervention group (n = 25) additionally received a home-based gait and balance training plan. Another sample of healthy adults (n = 32) was assessed once to analyze the impact of sensor placement (lower back vs. lower abdomen) on gait and balance analysis. Both the control and intervention group exhibited improvements in gait/stance. However, the trends over time were similar for both groups, suggesting that targeted training and repeated task performance equally contributed to the improvement of the measured variables. Since no significant differences were found in sensor placement, we suggest that a smartphone used as a wearable sensor could be worn both on the lower abdomen and the lower back in gait and balance analyses.

Predictive Data Analytics in Telecare and Telehealth: Systematic Scoping Review

BACKGROUND

Telecare and telehealth are important care-at-home services used to support individuals to live more independently at home. Historically, these technologies have reactively responded to issues. However, there has been a recent drive to make better use of the data from these services to facilitate more proactive and predictive care.

OBJECTIVE

This review seeks to explore the ways in which predictive data analytics techniques have been applied in telecare and telehealth in at-home settings.

METHODS

The PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) checklist was adhered to alongside Arksey and O'Malley's methodological framework. English language papers published in MEDLINE, Embase, and Social Science Premium Collection between 2012 and 2022 were considered and results were screened against inclusion or exclusion criteria.

RESULTS

In total, 86 papers were included in this review. The types of analytics featuring in this review can be categorized as anomaly detection (n=21), diagnosis (n=32), prediction (n=22), and activity recognition (n=11). The most common health conditions represented were Parkinson disease (n=12) and cardiovascular conditions (n=11). The main findings include: a lack of use of routinely collected data; a dominance of diagnostic tools; and barriers and opportunities that exist, such as including patient-reported outcomes, for future predictive analytics in telecare and telehealth.

CONCLUSIONS

All papers in this review were small-scale pilots and, as such, future research should seek to apply these predictive techniques into larger trials. Additionally, further integration of routinely collected care data and patient-reported outcomes into predictive models in telecare and telehealth offer significant opportunities to improve the analytics being performed and should be explored further. Data sets used must be of suitable size and diversity, ensuring that models are generalizable to a wider population and can be appropriately trained, validated, and tested.

Sensor-Based Quantification of MDS-UPDRS III Subitems in Parkinson's Disease Using Machine Learning

Wearable sensors could be beneficial for the continuous quantification of upper limb motor symptoms in people with Parkinson's disease (PD). This work evaluates the use of two inertial measurement units combined with supervised machine learning models to classify and predict a subset of MDS-UPDRS III subitems in PD. We attached the two compact wearable sensors on the dorsal part of each hand of 33 people with PD and 12 controls. Each participant performed six clinical movement tasks in parallel with an assessment of the MDS-UPDRS III. Random forest (RF) models were trained on the sensor data and motor scores. An overall accuracy of 94% was achieved in classifying the movement tasks. When employed for classifying the motor scores, the averaged area under the receiver operating characteristic values ranged from 68% to 92%. Motor scores were additionally predicted using an RF regression model. In a comparative analysis, trained support vector machine models outperformed the RF models for specific tasks. Furthermore, our results surpass the literature in certain cases. The methods developed in this work serve as a base for future studies, where home-based assessments of pharmacological effects on motor function could complement regular clinical assessments.

Data Augmentation Techniques for Accurate Action Classification in Stroke Patients with Hemiparesis

Stroke survivors with hemiparesis require extensive home-based rehabilitation. Deep learning-based classifiers can detect actions and provide feedback based on patient data; however, this is difficult owing to data sparsity and heterogeneity. In this study, we investigate data augmentation and model training strategies to address this problem. Three transformations are tested with varying data volumes to analyze the changes in the classification performance of individual data. Moreover, the impact of transfer learning relative to a pre-trained one-dimensional convolutional neural network (Conv1D) and training with an advanced InceptionTime model are estimated with data augmentation. In Conv1D, the joint training data of non-disabled (ND) participants and double rotationally augmented data of stroke patients is observed to outperform the baseline in terms of F1-score (60.9% vs. 47.3%). Transfer learning pre-trained with ND data exhibits 60.3% accuracy, whereas joint training with InceptionTime exhibits 67.2% accuracy under the same conditions. Our results indicate that rotational augmentation is more effective for individual data with initially lower performance and subset data with smaller numbers of participants than other techniques, suggesting that joint training on rotationally augmented ND and stroke data enhances classification performance, particularly in cases with sparse data and lower initial performance.

Validation of Proprietary and Novel Step-counting Algorithms for Individuals Ambulating With a Lower Limb Prosthesis

OBJECTIVE

To compare the accuracy and reliability of 10 different accelerometer-based step-counting algorithms for individuals with lower limb loss, accounting for different clinical characteristics and real-world activities.

DESIGN

Cross-sectional study.

SETTING

General community setting (ie, institutional research laboratory and community free-living).

PARTICIPANTS

Forty-eight individuals with a lower limb amputation (N=48) wore an ActiGraph (AG) wGT3x-BT accelerometer proximal to the foot of their prosthetic limb during labeled indoor/outdoor activities and community free-living.

INTERVENTIONS

Not applicable.

MAIN OUTCOME MEASURES

Intraclass correlation coefficient (ICC), absolute and root mean square error (RMSE), and Bland Altman plots were used to compare true (manual) step counts to estimated step counts from the proprietary AG Default algorithm and low frequency extension filter, as well as from 8 novel algorithms based on continuous wavelet transforms, fast Fourier transforms (FFTs), and peak detection.

RESULTS

All algorithms had excellent agreement with manual step counts (ICC>0.9). The AG Default and FFT algorithms had the highest overall error (RMSE=17.81 and 19.91 steps, respectively), widest limits of agreement, and highest error during outdoor and ramp ambulation. The AG Default algorithm also had among the highest error during indoor ambulation and stairs, while a FFT algorithm had the highest error during stationary tasks. Peak detection algorithms, especially those using pre-set parameters with a trial-specific component, had among the lowest error across all activities (RMSE=4.07-8.99 steps).

CONCLUSIONS

Because of its simplicity and accuracy across activities and clinical characteristics, we recommend the peak detection algorithm with set parameters to count steps using a prosthetic-worn AG among individuals with lower limb loss for clinical and research applications.

Investigating Activity Recognition for Hemiparetic Stroke Patients Using Wearable Sensors: A Deep Learning Approach with Data Augmentation

Measuring the daily use of an affected limb after hospital discharge is crucial for hemiparetic stroke rehabilitation. Classifying movements using non-intrusive wearable sensors provides context for arm use and is essential for the development of a home rehabilitation system. However, the movement classification of stroke patients poses unique challenges, including variability and sparsity. To address these challenges, we collected movement data from 15 hemiparetic stroke patients (Stroke group) and 29 non-disabled individuals (ND group). The participants performed two different tasks, the range of motion (14 movements) task and the activities of daily living (56 movements) task, wearing five inertial measurement units in a home setting. We trained a 1D convolutional neural network and evaluated its performance for different training groups: ND-only, Stroke-only, and ND and Stroke jointly. We further compared the model performance with data augmentation from axis rotation and investigated how the performance varied based on the asymmetry of movements. The joint training of ND + Stroke yielded an increased F1-score by a margin of 31.6% and 10.6% compared to ND-only training and Stroke-only training, respectively. Data augmentation further enhanced F1-scores across all conditions by an average of 11.3%. Finally, asymmetric movements decreased the F1-score by 25.9% compared to symmetric movements in the Stroke group, indicating the importance of asymmetry in movement classification.

AI in Rehabilitation Medicine: Opportunities and Challenges

Artificial intelligence (AI) tools are increasingly able to learn from larger and more complex data, thus allowing clinicians and scientists to gain new insights from the information they collect about their patients every day. In rehabilitation medicine, AI can be used to find patterns in huge amounts of healthcare data. These patterns can then be leveraged at the individual level, to design personalized care strategies and interventions to optimize each patient's outcomes. However, building effective AI tools requires many careful considerations about how we collect and handle data, how we train the models, and how we interpret results. In this perspective, we discuss some of the current opportunities and challenges for AI in rehabilitation. We first review recent trends in AI for the screening, diagnosis, treatment, and continuous monitoring of disease or injury, with a special focus on the different types of healthcare data used for these applications. We then examine potential barriers to designing and integrating AI into the clinical workflow, and we propose an end-to-end framework to address these barriers and guide the development of effective AI for rehabilitation. Finally, we present ideas for future work to pave the way for AI implementation in real-world rehabilitation practices.

A scoping review on recent trends in wearable sensors to analyze gait in people with stroke: From sensor placement to validation against gold-standard equipment

The purpose of the review is to evaluate wearable sensor placement, their impact and validation of wearable sensors on analyzing gait, primarily the postural instability in people with stroke. Databases, namely PubMed, Cochrane, SpringerLink, and IEEE Xplore were searched to identify related articles published since January 2005. The authors have selected the articles by considering patient characteristics, intervention details, and outcome measurements by following the priorly set inclusion and exclusion criteria. From a total of 1077 articles, 142 were included in this study and classified into functional fields, namely postural stability (PS) assessments, physical activity monitoring (PA), gait pattern classification (GPC), and foot drop correction (FDC). The review covers the types of wearable sensors, their placement, and their performance in terms of reliability and validity. When employing a single wearable sensor, the pelvis and foot were the most used locations for detecting gait asymmetry and kinetic parameters, respectively. Multiple Inertial Measurement Units placed at different body parts were effectively used to estimate postural stability and gait pattern. This review article has compared results of placement of sensors at different locations helping researchers and clinicians to identify the best possible placement for sensors to measure specific kinematic and kinetic parameters in persons with stroke.

Improving Inertial Sensor-Based Activity Recognition in Neurological Populations

Inertial sensor-based human activity recognition (HAR) has a range of healthcare applications as it can indicate the overall health status or functional capabilities of people with impaired mobility. Typically, artificial intelligence models achieve high recognition accuracies when trained with rich and diverse inertial datasets. However, obtaining such datasets may not be feasible in neurological populations due to, e.g., impaired patient mobility to perform many daily activities. This study proposes a novel framework to overcome the challenge of creating rich and diverse datasets for HAR in neurological populations. The framework produces images from numerical inertial time-series data (initial state) and then artificially augments the number of produced images (enhanced state) to achieve a larger dataset. Here, we used convolutional neural network (CNN) architectures by utilizing image input. In addition, CNN enables transfer learning which enables limited datasets to benefit from models that are trained with big data. Initially, two benchmarked public datasets were used to verify the framework. Afterward, the approach was tested in limited local datasets of healthy subjects (HS), Parkinson's disease (PD) population, and stroke survivors (SS) to further investigate validity. The experimental results show that when data augmentation is applied, recognition accuracies have been increased in HS, SS, and PD by 25.6%, 21.4%, and 5.8%, respectively, compared to the no data augmentation state. In addition, data augmentation contributes to better detection of stair ascent and stair descent by 39.1% and 18.0%, respectively, in limited local datasets. Findings also suggest that CNN architectures that have a small number of deep layers can achieve high accuracy. The implication of this study has the potential to reduce the burden on participants and researchers where limited datasets are accrued.

Accuracy of gait and posture classification using movement sensors in individuals with mobility impairment after stroke

Background: Stroke leads to motor impairment which reduces physical activity, negatively affects social participation, and increases the risk of secondary cardiovascular events. Continuous monitoring of physical activity with motion sensors is promising to allow the prescription of tailored treatments in a timely manner. Accurate classification of gait activities and body posture is necessary to extract actionable information for outcome measures from unstructured motion data. We here develop and validate a solution for various sensor configurations specifically for a stroke population.

Methods: Video and movement sensor data (locations: wrists, ankles, and chest) were collected from fourteen stroke survivors with motor impairment who performed real-life activities in their home environment. Video data were labeled for five classes of gait and body postures and three classes of transitions that served as ground truth. We trained support vector machine (SVM), logistic regression (LR), and k-nearest neighbor (kNN) models to identify gait bouts only or gait and posture. Model performance was assessed by the nested leave-one-subject-out protocol and compared across five different sensor placement configurations.

Results: Our method achieved very good performance when predicting real-life gait versus non-gait (Gait classification) with an accuracy between 85% and 93% across sensor configurations, using SVM and LR modeling. On the much more challenging task of discriminating between the body postures lying, sitting, and standing as well as walking, and stair ascent/descent (Gait and postures classification), our method achieves accuracies between 80% and 86% with at least one ankle and wrist sensor attached unilaterally. The Gait and postures classification performance between SVM and LR was equivalent but superior to kNN.

Conclusion: This work presents a comparison of performance when classifying Gait and body postures in post-stroke individuals with different sensor configurations, which provide options for subsequent outcome evaluation. We achieved accurate classification of gait and postures performed in a real-life setting by individuals with a wide range of motor impairments due to stroke. This validated classifier will hopefully prove a useful resource to researchers and clinicians in the increasingly important field of digital health in the form of remote movement monitoring using motion sensors.

Facilitators and barriers to real-life mobility in community-dwelling older adults: a narrative review of accelerometry- and global positioning system-based studies

Real-life mobility, also called "enacted" mobility, characterizes an individual's activity and participation in the community. Real-life mobility may be facilitated or hindered by a variety of factors, such as physical abilities, cognitive function, psychosocial aspects, and external environment characteristics. Advances in technology have allowed for objective quantification of real-life mobility using wearable sensors, specifically, accelerometry and global positioning systems (GPSs). In this review article, first, we summarize the common mobility measures extracted from accelerometry and GPS. Second, we summarize studies assessing the associations of facilitators and barriers influencing mobility of community-dwelling older adults with mobility measures from sensor technology. We found the most used accelerometry measures focus on the duration and intensity of activity in daily life. Gait quality measures, e.g., cadence, variability, and symmetry, are not usually included. GPS has been used to investigate mobility behavior, such as spatial and temporal measures of path traveled, location nodes traversed, and mode of transportation. Factors of note that facilitate/hinder community mobility were cognition and psychosocial influences. Fewer studies have included the influence of external environments, such as sidewalk quality, and socio-economic status in defining enacted mobility. Increasing our understanding of the facilitators and barriers to enacted mobility can inform wearable technology-enabled interventions targeted at delaying mobility-related disability and improving participation of older adults in the community.

Use of artificial intelligence as an instrument of evaluation after stroke: a scoping review based on international classification of functioning, disability and health concept

INTRODUCTION

To understand the current practices in stroke evaluation, the main clinical decision support system and artificial intelligence (AI) technologies need to be understood to assist the therapist in obtaining better insights about impairments and level of activity and participation in persons with stroke during rehabilitation.

METHODS

This scoping review maps the use of AI for the functional evaluation of persons with stroke; the context involves any setting of rehabilitation. Data were extracted from CENTRAL, MEDLINE, EMBASE, LILACS, CINAHL, PEDRO Web of Science, IEEE Xplore, AAAI Publications, ACM Digital Library, MathSciNet, and arXiv up to January 2021. The data obtained from the literature review were summarized in a single dataset in which each reference paper was considered as an instance, and the study characteristics were considered as attributes. The attributes used for the multiple correspondence analysis were publication year, study type, sample size, age, stroke phase, stroke type, functional status, AI type, and AI function.

RESULTS

Forty-four studies were included. The analysis showed that spasticity analysis based on ML techniques was used for the cases of stroke with moderate functional status. The techniques of deep learning and pressure sensors were used for gait analysis. Machine learning techniques and algorithms were used for upper limb and reaching analyses. The inertial measurement unit technique was applied in studies where the functional status was between mild and severe. The fuzzy logic technique was used for activity classifiers.

CONCLUSION

The prevailing research themes demonstrated the growing utility of AI algorithms for stroke evaluation.

Wearable airbag technology and machine learned models to mitigate falls after stroke

BACKGROUND

Falls are a common complication experienced after a stroke and can cause serious detriments to physical health and social mobility, necessitating a dire need for intervention. Among recent advancements, wearable airbag technology has been designed to detect and mitigate fall impact. However, these devices have not been designed nor validated for the stroke population and thus, may inadequately detect falls in individuals with stroke-related motor impairments. To address this gap, we investigated whether population-specific training data and modeling parameters are required to pre-detect falls in a chronic stroke population.

METHODS

We collected data from a wearable airbag's inertial measurement units (IMUs) from individuals with (n = 20 stroke) and without (n = 15 control) history of stroke while performing a series of falls (842 falls total) and non-falls (961 non-falls total) in a laboratory setting. A leave-one-subject-out crossvalidation was used to compare the performance of two identical machine learned models (adaptive boosting classifier) trained on cohort-dependent data (control or stroke) to pre-detect falls in the stroke cohort.

RESULTS

The average performance of the model trained on stroke data (recall = 0.905, precision = 0.900) had statistically significantly better recall (P = 0.0035) than the model trained on control data (recall = 0.800, precision = 0.944), while precision was not statistically significantly different. Stratifying models trained on specific fall types revealed differences in pre-detecting anterior-posterior (AP) falls (stroke-trained model's F₁-score was 35% higher, P = 0.019). Using activities of daily living as non-falls training data (compared to near-falls) significantly increased the AUC (Area under the receiver operating characteristic) for classifying AP falls for both models (P < 0.04). Preliminary analysis suggests that users with more severe stroke impairments benefit further from a stroke-trained model. The optimal lead time (time interval pre-impact to detect falls) differed between control- and stroke-trained models.

CONCLUSIONS

These results demonstrate the importance of population sensitivity, non-falls data, and optimal lead time for machine learned pre-impact fall detection specific to stroke. Existing fall mitigation technologies should be challenged to include data of neurologically impaired individuals in model development to adequately detect falls in other high fall risk populations. Trial registration https://clinicaltrials.gov/ct2/show/NCT05076565 ; Unique Identifier: NCT05076565. Retrospectively registered on 13 October 2021.

Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review

Recent years have seen significant advances in the sensing capabilities of smartphones, enabling them to collect rich contextual information such as location, device usage, and human activity at a given point in time. Combined with widespread user adoption and the ability to gather user data remotely, smartphone-based sensing has become an appealing choice for health research. Numerous studies over the years have demonstrated the promise of using smartphone-based sensing to monitor a range of health conditions, particularly mental health conditions. However, as research is progressing to develop the predictive capabilities of smartphones, it becomes even more crucial to fully understand the capabilities and limitations of using this technology, given its potential impact on human health. To this end, this paper presents a narrative review of smartphone-sensing literature from the past 5 years, to highlight the opportunities and challenges of this approach in healthcare. It provides an overview of the type of health conditions studied, the types of data collected, tools used, and the challenges encountered in using smartphones for healthcare studies, which aims to serve as a guide for researchers wishing to embark on similar research in the future. Our findings highlight the predominance of mental health studies, discuss the opportunities of using standardized sensing approaches and machine-learning advancements, and present the trends of smartphone sensing in healthcare over the years.

Foot-to-Ground Phases Detection: A Comparison of Data Representation Formatting Methods with Respect to Adaption of Deep Learning Architectures

Identifying the foot stance and foot swing phases, also known as foot-to-ground (FTG) detection, is a branch of Human Activity Recognition (HAR). Our study aims to detect two main phases of the gait (i.e., foot-off and foot-contact) corresponding to the moments when each foot is in contact with the ground or not. This will allow the medical professionals to characterize and identify the different phases of the human gait and their respective patterns. This detection process is paramount for extracting gait features (e.g., step width, stride width, gait speed, cadence, etc.) used by medical experts to highlight gait anomalies, stance issues, or any other walking irregularities. It will be used to assist health practitioners with patient monitoring, in addition to developing a full pipeline for FTG detection that would help compute gait indicators. In this paper, a comparison of different training configurations, including model architectures, data formatting, and pre-processing, was conducted to select the parameters leading to the highest detection accuracy. This binary classification provides a label for each timestamp informing whether the foot is in contact with the ground or not. Models such as CNN, LSTM, and ConvLSTM were the best fits for this study. Yet, we did not exclude DNNs and Machine Learning models, such as Random Forest and XGBoost from our work in order to have a wide range of possible comparisons. As a result of our experiments, which included 27 senior participants who had a stroke in the past wearing IMU sensors on their ankles, the ConvLSTM model achieved a high accuracy of 97.01% for raw windowed data with a size of 3 frames per window, and each window was formatted to have two superimposed channels (accelerometer and gyroscope channels). The model was trained to have the best detection without any knowledge of the participants’ personal information including age, gender, health condition, the type of activity, or the used foot. In other words, the model’s input data only originated from IMU sensors. Overall, in terms of FTG detection, the combination of the ConvLSTM model and the data representation had an important impact in outperforming other start-of-the-art configurations; in addition, the compromise between the model’s complexity and its accuracy is a major asset for deploying this model and developing real-time solutions.

Accuracy comparison of machine learning algorithms at various wear-locations for activity identification post stroke: A pilot analysis. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2021;2021:6106-6109. [PMID: 34892510 DOI: 10.1109/embc46164.2021.9630745] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Objective and accurate activity identification of physical activities in everyday life is an important aspect in assessing the impact of various post-stroke rehabilitation therapies and interventions. Since post-stroke hemiparesis affects gait and balance in individuals with stroke, activity identification algorithms that consider stroke-specific movement irregularities are needed. While wearable physical activity monitors provide the means to detect activities in the free-living, algorithms using their data are specific to the wear location of the device. This pilot study builds, validates, and compares three machine learning algorithms (linear support vector machine, Random Forest, and RUSBoosted trees) at three popular wear locations (wrist, waist, and ankle) to identify and accurately distinguish mobility-related activities (sitting, standing and walking) in individuals with chronic stroke. A total of 102 minutes of data from two lab visits of three-stroke participants was used to build the classifiers. A 5-fold cross-validation technique was used to validate and compare the accuracy of classifiers. RUSBoosted trees using data from waist and ankle activity monitors, with an accuracy of 99.1%, outperformed other classifiers in detecting three activities of interest.Clinical Relevance- One of the major aims of post-stroke rehabilitation is improving mobility, which may be facilitated by understanding the structure and pattern of everyday mobility through real-world, objective outcomes. Accurate activity identification, as shown in this pilot investigation, is an essential first step before developing objective outcomes for monitoring mobility and balance in everyday life of these individuals.

Wearable devices for tracking physical activity in the community after an acquired brain injury: A systematic review

OBJECTIVE

The application of wearable devices in individuals with acquired brain injury (ABI) resulting from stroke or traumatic brain injury (TBI) for monitoring physical activity (PA) has been relatively recent. The current systematic review aims to provide insights into the adaption of these devices, outcome metrics, and their transition from the laboratory to the community for PA monitoring of individuals with ABI.

LITERATURE SURVEY

The PubMed and Google Scholar databases were systematically reviewed using appropriate search terms. A total of 20 articles were reviewed from the past 15 years.

METHODOLOGY

Articles were classified into three categories - PA measurement studies, PA classification studies, and validation studies. The quality of studies was assessed using a quality appraisal checklist.

SYNTHESIS

It was found that the transition of wearable devices from in-lab to community-based studies in individuals with stroke has started but is not widespread. The transition of wearable devices in the community has not yet started for individuals with TBI. Accelerometer-based devices were more frequently chosen than pedometers and inertial measurement units. No consensus on a preferred wearable device (make or model) or wear location could be identified, though step count was the most common outcome metric. The accuracy and validity of most outcome metrics used in the community were not reported for many studies.

CONCLUSIONS

To facilitate future studies use wearable devices for PA measurement in the community, we recommend that researchers provide details on the accuracy and validity of the outcome metrics specific to the study environment. Once the accuracy and validity are established for a specific population, wearable devices and their derived outcomes can provide objective information on mobility impairment as well as the effect of rehabilitation in the community. This article is protected by copyright. All rights reserved.

Upper-Limb Motion Recognition Based on Hybrid Feature Selection: Algorithm Development and Validation

Background

For rehabilitation training systems, it is essential to automatically record and recognize exercises, especially when more than one type of exercise is performed without a predefined sequence. Most motion recognition methods are based on feature engineering and machine learning algorithms. Time-domain and frequency-domain features are extracted from original time series data collected by sensor nodes. For high-dimensional data, feature selection plays an important role in improving the performance of motion recognition. Existing feature selection methods can be categorized into filter and wrapper methods. Wrapper methods usually achieve better performance than filter methods; however, in most cases, they are computationally intensive, and the feature subset obtained is usually optimized only for the specific learning algorithm.

Objective

This study aimed to provide a feature selection method for motion recognition of upper-limb exercises and improve the recognition performance.

Methods

Motion data from 5 types of upper-limb exercises performed by 21 participants were collected by a customized inertial measurement unit (IMU) node. A total of 60 time-domain and frequency-domain features were extracted from the original sensor data. A hybrid feature selection method by combining filter and wrapper methods (FESCOM) was proposed to eliminate irrelevant features for motion recognition of upper-limb exercises. In the filter stage, candidate features were first selected from the original feature set according to the significance for motion recognition. In the wrapper stage, k-nearest neighbors (kNN), Naïve Bayes (NB), and random forest (RF) were evaluated as the wrapping components to further refine the features from the candidate feature set. The performance of the proposed FESCOM method was verified using experiments on motion recognition of upper-limb exercises and compared with the traditional wrapper method.

Results

Using kNN, NB, and RF as the wrapping components, the classification error rates of the proposed FESCOM method were 1.7%, 8.9%, and 7.4%, respectively, and the feature selection time in each iteration was 13 seconds, 71 seconds, and 541 seconds, respectively.

Conclusions

The experimental results demonstrated that, in the case of 5 motion types performed by 21 healthy participants, the proposed FESCOM method using kNN and NB as the wrapping components achieved better recognition performance than the traditional wrapper method. The FESCOM method dramatically reduces the search time in the feature selection process. The results also demonstrated that the optimal number of features depends on the classifier. This approach serves to improve feature selection and classification algorithm selection for upper-limb motion recognition based on wearable sensor data, which can be extended to motion recognition of more motion types and participants.

Combining Accelerometer and GPS Features to Evaluate Community Mobility in Knee Ankle Foot Orthoses (KAFO) Users

Orthotic and assistive devices such as knee ankle foot orthoses (KAFO), come in a variety of forms and fits, with several levels of available features that could help users perform daily activities more naturally. However, objective data on the actual use of these devices outside of the research lab is usually not obtained. Such data could enhance traditional lab-based outcome measures and inform clinical decision-making when prescribing new orthotic and assistive technology. Here, we link data from a GPS unit and an accelerometer mounted on the orthotic device to quantify its usage in the community and examine the correlations with clinical metrics. We collected data from 14 individuals over a period of 2 months as they used their personal KAFO first, and then a novel research KAFO; for each device we quantified number of steps, cadence, time spent at community locations and time wearing the KAFO at those locations. Sensor-derived metrics showed that mobility patterns differed widely between participants (mean steps: 591.3, SD =704.2). The novel KAFO generally enabled participants to walk faster during clinical tests ( ∆6 Minute-Walk-Test=71.5m, p=0.006). However, some participants wore the novel device less often despite improved performance on these clinical measures, leading to poor correlation between changes in clinical outcome measures and changes in community mobility ( ∆6 Minute-Walk-Test - ∆ Community Steps: r=0.09, p=0.76). Our results suggest that some traditional clinical outcome measures may not be associated with the actual wear time of an assistive device in the community, and obtaining personalized data from real-world use through wearable technology is valuable.

Systematic review on the application of wearable inertial sensors to quantify everyday life motor activity in people with mobility impairments

BACKGROUND

Recent advances in wearable sensor technologies enable objective and long-term monitoring of motor activities in a patient's habitual environment. People with mobility impairments require appropriate data processing algorithms that deal with their altered movement patterns and determine clinically meaningful outcome measures. Over the years, a large variety of algorithms have been published and this review provides an overview of their outcome measures, the concepts of the algorithms, the type and placement of required sensors as well as the investigated patient populations and measurement properties.

METHODS

A systematic search was conducted in MEDLINE, EMBASE, and SCOPUS in October 2019. The search strategy was designed to identify studies that (1) involved people with mobility impairments, (2) used wearable inertial sensors, (3) provided a description of the underlying algorithm, and (4) quantified an aspect of everyday life motor activity. The two review authors independently screened the search hits for eligibility and conducted the data extraction for the narrative review.

RESULTS

Ninety-five studies were included in this review. They covered a large variety of outcome measures and algorithms which can be grouped into four categories: (1) maintaining and changing a body position, (2) walking and moving, (3) moving around using a wheelchair, and (4) activities that involve the upper extremity. The validity or reproducibility of these outcomes measures was investigated in fourteen different patient populations. Most of the studies evaluated the algorithm's accuracy to detect certain activities in unlabeled raw data. The type and placement of required sensor technologies depends on the activity and outcome measure and are thoroughly described in this review. The usability of the applied sensor setups was rarely reported.

CONCLUSION

This systematic review provides a comprehensive overview of applications of wearable inertial sensors to quantify everyday life motor activity in people with mobility impairments. It summarizes the state-of-the-art, it provides quick access to the relevant literature, and it enables the identification of gaps for the evaluation of existing and the development of new algorithms.

Using an Accelerometer-Based Step Counter in Post-Stroke Patients: Validation of a Low-Cost Tool

Monitoring the real-life mobility of stroke patients could be extremely useful for clinicians. Step counters are a widely accessible, portable, and cheap technology that can be used to monitor patients in different environments. The aim of this study was to validate a low-cost commercial tri-axial accelerometer-based step counter for stroke patients and to determine the best positioning of the step counter (wrists, ankles, and waist). Ten healthy subjects and 43 post-stroke patients were enrolled and performed four validated clinical tests (10 m, 50 m, and 6 min walking tests and timed up and go tests) while wearing five step counters in different positions while a trained operator counted the number of steps executed in each test manually. Data from step counters and those collected manually were compared using the intraclass coefficient correlation and mean average percentage error. The Bland–Altman plot was also used to describe agreement between the two quantitative measurements (step counter vs. manual counting). During walking tests in healthy subjects, the best reliability was found for lower limbs and waist placement (intraclass coefficient correlations (ICCs) from 0.46 to 0.99), and weak reliability was observed for upper limb placement in every test (ICCs from 0.06 to 0.38). On the contrary, in post-stroke patients, moderate reliability was found only for the lower limbs in the 6 min walking test (healthy ankle ICC: 0.69; pathological ankle ICC: 0.70). Furthermore, the Bland–Altman plot highlighted large average discrepancies between methods for the pathological group. However, while the step counter was not able to reliably determine steps for slow patients, when applied to the healthy ankle of patients who walked faster than 0.8 m/s, it counted steps with excellent precision, similar to that seen in the healthy subjects (ICCs from 0.36 to 0.99). These findings show that a low-cost accelerometer-based step counter could be useful for measuring mobility in select high-performance patients and could be used in clinical and real-world settings.

Role of data measurement characteristics in the accurate detection of Parkinson's disease symptoms using wearable sensors

Background

Parkinson’s disease (PD) is a progressive neurological disease, with characteristic motor symptoms such as tremor and bradykinesia. There is a growing interest to continuously monitor these and other symptoms through body-worn sensor technology. However, limited battery life and memory capacity hinder the potential for continuous, long-term monitoring with these devices. There is little information available on the relative value of adding sensors, increasing sampling rate, or computing complex signal features, all of which may improve accuracy of symptom detection at the expense of computational resources. Here we build on a previous study to investigate the relationship between data measurement characteristics and accuracy when using wearable sensor data to classify tremor and bradykinesia in patients with PD.

Methods

Thirteen individuals with PD wore a flexible, skin-mounted sensor (collecting tri-axial accelerometer and gyroscope data) and a commercial smart watch (collecting tri-axial accelerometer data) on their predominantly affected hand. The participants performed a series of standardized motor tasks, during which a clinician scored the severity of tremor and bradykinesia in that limb. Machine learning models were trained on scored data to classify tremor and bradykinesia. Model performance was compared when using different types of sensors (accelerometer and/or gyroscope), different data sampling rates (up to 62.5 Hz), and different categories of pre-engineered features (up to 148 features). Performance was also compared between the flexible sensor and smart watch for each analysis.

Results

First, there was no effect of device type for classifying tremor symptoms (p > 0.34), but bradykinesia models incorporating gyroscope data performed slightly better (up to 0.05 AUROC) than other models (p = 0.01). Second, model performance decreased with sampling frequency (p < 0.001) for tremor, but not bradykinesia (p > 0.47). Finally, model performance for both symptoms was maintained after substantially reducing the feature set.

Conclusions

Our findings demonstrate the ability to simplify measurement characteristics from body-worn sensors while maintaining performance in PD symptom detection. Understanding the trade-off between model performance and data resolution is crucial to design efficient, accurate wearable sensing systems. This approach may improve the feasibility of long-term, continuous, and real-time monitoring of PD symptoms by reducing computational burden on wearable devices.

A review of computational approaches for evaluation of rehabilitation exercises

Recent advances in data analytics and computer-aided diagnostics stimulate the vision of patient-centric precision healthcare, where treatment plans are customized based on the health records and needs of every patient. In physical rehabilitation, the progress in machine learning and the advent of affordable and reliable motion capture sensors have been conducive to the development of approaches for automated assessment of patient performance and progress toward functional recovery. The presented study reviews computational approaches for evaluating patient performance in rehabilitation programs using motion capture systems. Such approaches will play an important role in supplementing traditional rehabilitation assessment performed by trained clinicians, and in assisting patients participating in home-based rehabilitation. The reviewed computational methods for exercise evaluation are grouped into three main categories: discrete movement score, rule-based, and template-based approaches. The review places an emphasis on the application of machine learning methods for movement evaluation in rehabilitation. Related work in the literature on data representation, feature engineering, movement segmentation, and scoring functions is presented. The study also reviews existing sensors for capturing rehabilitation movements and provides an informative listing of pertinent benchmark datasets. The significance of this paper is in being the first to provide a comprehensive review of computational methods for evaluation of patient performance in rehabilitation programs.

Hidden Markov model-based activity recognition for toddlers

OBJECTIVE

Physical activity has been shown to impact future health outcomes in adults, but little is known about the long-term impact of physical activity in toddlers. Accurately measuring the specific types and amounts of physical activity in toddlers will help us to understand, predict, and better affect their future health outcomes. Although activity recognition has been extensively developed for adults as well as older children, toddlers move in ways that are significantly different from older children, indicating the need for a more tailored approach.

APPROACH

In this study, 22 toddlers wore Actigraph waist-worn accelerometers which recorded their movements during guided play. The toddlers were videotaped and their activities were later annotated for the following eight distinct activity classes: lying down, being carried, riding in a stroller, sitting, standing, running/walking, crawling, and climbing up/down. Accelerometer data were extracted in 2 s signal windows and paired with the activities the toddlers were performing during that time interval.

MAIN RESULTS

A variety of classifiers were tuned to a validation set. A random forest classifier was found to achieve the highest accuracy of 63.8% in a test set. To improve the accuracy, a hidden Markov model (HMM) was applied by providing the predictions of the static classifiers as observations. The HMM was able to improve the accuracy to 64.8% with all five classifiers increasing the accuracy an average of 1.3% points (95% confidence interval  =  0.7-1.9, p   <  0.01). When the three most misclassified activities (sitting, standing, and riding in a stroller) were collapsed together, the accuracy increased to 79.3%.

SIGNIFICANCE

Further refinement of the toddler activity recognition classifier will enable more accurate measurements of toddler activity and improve future health outcomes of toddlers.

A Stroke Risk Detection: Improving Hybrid Feature Selection Method

Background

Stroke is one of the most common diseases that cause mortality. Detecting the risk of stroke for individuals is critical yet challenging because of a large number of risk factors for stroke.

Objective

This study aimed to address the limitation of ineffective feature selection in existing research on stroke risk detection. We have proposed a new feature selection method called weighting- and ranking-based hybrid feature selection (WRHFS) to select important risk factors for detecting ischemic stroke.

Methods

WRHFS integrates the strengths of various filter algorithms by following the principle of a wrapper approach. We employed a variety of filter-based feature selection models as the candidate set, including standard deviation, Pearson correlation coefficient, Fisher score, information gain, Relief algorithm, and chi-square test and used sensitivity, specificity, accuracy, and Youden index as performance metrics to evaluate the proposed method.

Results

This study chose 792 samples from the electronic records of 13,421 patients in a community hospital. Each sample included 28 features (24 blood test features and 4 demographic features). The results of evaluation showed that the proposed method selected 9 important features out of the original 28 features and significantly outperformed baseline methods. Their cumulative contribution was 0.51. The WRHFS method achieved a sensitivity of 82.7% (329/398), specificity of 80.4% (317/394), classification accuracy of 81.5% (645/792), and Youden index of 0.63 using only the top 9 features. We have also presented a chart for visualizing the risk of having ischemic strokes.

Conclusions

This study has proposed, developed, and evaluated a new feature selection method for identifying the most important features for building effective and parsimonious models for stroke risk detection. The findings of this research provide several novel research contributions and practical implications.

Smartwatches Can Detect Walker and Cane Use in Older Adults

BACKGROUND AND OBJECTIVES

Clinicians commonly prescribe assistive devices such as walkers or canes to reduce older adults' fall risk. However, older adults may not consistently use their assistive device, and measuring adherence can be challenging due to self-report bias or cognitive deficits. Because walking patterns can change while using an assistive device, we hypothesized that smartphones and smartwatches, combined with machine-learning algorithms, could detect whether an older adult was walking with an assistive device.

RESEARCH DESIGN AND METHODS

Older adults at an Adult Day Center (n = 14) wore an Android smartphone and Actigraph smartwatch while completing the six-minute walk, 10-meter walk, and Timed Up and Go tests with and without their assistive device on five separate days. We used accelerometer data from the devices to build machine-learning algorithms to detect whether the participant was walking with or without their assistive device. We tested our algorithms using cross-validation.

RESULTS

Smartwatch classifiers could accurately detect assistive device use, but smartphone classifiers performed poorly. Customized smartwatch classifiers, which were created specifically for one participant, had greater than 95% classification accuracy for all participants. Noncustomized smartwatch classifiers (ie, an "off-the-shelf" system) had greater than 90% accuracy for 10 of the 14 participants. A noncustomized system performed better for walker users than cane users.

DISCUSSION AND IMPLICATIONS

Our approach can leverage data from existing commercial devices to provide a deeper understanding of walker or cane use. This work can inform scalable public health monitoring tools to quantify assistive device adherence and enable proactive fall interventions.

Machine learning in human movement biomechanics: Best practices, common pitfalls, and new opportunities

Traditional laboratory experiments, rehabilitation clinics, and wearable sensors offer biomechanists a wealth of data on healthy and pathological movement. To harness the power of these data and make research more efficient, modern machine learning techniques are starting to complement traditional statistical tools. This survey summarizes the current usage of machine learning methods in human movement biomechanics and highlights best practices that will enable critical evaluation of the literature. We carried out a PubMed/Medline database search for original research articles that used machine learning to study movement biomechanics in patients with musculoskeletal and neuromuscular diseases. Most studies that met our inclusion criteria focused on classifying pathological movement, predicting risk of developing a disease, estimating the effect of an intervention, or automatically recognizing activities to facilitate out-of-clinic patient monitoring. We found that research studies build and evaluate models inconsistently, which motivated our discussion of best practices. We provide recommendations for training and evaluating machine learning models and discuss the potential of several underutilized approaches, such as deep learning, to generate new knowledge about human movement. We believe that cross-training biomechanists in data science and a cultural shift toward sharing of data and tools are essential to maximize the impact of biomechanics research.

Wearable Movement Sensors for Rehabilitation: A Focused Review of Technological and Clinical Advances

Recent technologic advancements have enabled the creation of portable, low-cost, and unobtrusive sensors with tremendous potential to alter the clinical practice of rehabilitation. The application of wearable sensors to track movement has emerged as a promising paradigm to enhance the care provided to patients with neurologic or musculoskeletal conditions. These sensors enable quantification of motor behavior across disparate patient populations and emerging research shows their potential for identifying motor biomarkers, differentiating between restitution and compensation motor recovery mechanisms, remote monitoring, telerehabilitation, and robotics. Moreover, the big data recorded across these applications serve as a pathway to personalized and precision medicine. This article presents state-of-the-art and next-generation wearable movement sensors, ranging from inertial measurement units to soft sensors. An overview of clinical applications is presented across a wide spectrum of conditions that have potential to benefit from wearable sensors, including stroke, movement disorders, knee osteoarthritis, and running injuries. Complementary applications enabled by next-generation sensors that will enable point-of-care monitoring of neural activity and muscle dynamics during movement also are discussed.

How a diverse research ecosystem has generated new rehabilitation technologies: Review of NIDILRR's Rehabilitation Engineering Research Centers

Over 50 million United States citizens (1 in 6 people in the US) have a developmental, acquired, or degenerative disability. The average US citizen can expect to live 20% of his or her life with a disability. Rehabilitation technologies play a major role in improving the quality of life for people with a disability, yet widespread and highly challenging needs remain. Within the US, a major effort aimed at the creation and evaluation of rehabilitation technology has been the Rehabilitation Engineering Research Centers (RERCs) sponsored by the National Institute on Disability, Independent Living, and Rehabilitation Research. As envisioned at their conception by a panel of the National Academy of Science in 1970, these centers were intended to take a "total approach to rehabilitation", combining medicine, engineering, and related science, to improve the quality of life of individuals with a disability. Here, we review the scope, achievements, and ongoing projects of an unbiased sample of 19 currently active or recently terminated RERCs. Specifically, for each center, we briefly explain the needs it targets, summarize key historical advances, identify emerging innovations, and consider future directions. Our assessment from this review is that the RERC program indeed involves a multidisciplinary approach, with 36 professional fields involved, although 70% of research and development staff are in engineering fields, 23% in clinical fields, and only 7% in basic science fields; significantly, 11% of the professional staff have a disability related to their research. We observe that the RERC program has substantially diversified the scope of its work since the 1970's, addressing more types of disabilities using more technologies, and, in particular, often now focusing on information technologies. RERC work also now often views users as integrated into an interdependent society through technologies that both people with and without disabilities co-use (such as the internet, wireless communication, and architecture). In addition, RERC research has evolved to view users as able at improving outcomes through learning, exercise, and plasticity (rather than being static), which can be optimally timed. We provide examples of rehabilitation technology innovation produced by the RERCs that illustrate this increasingly diversifying scope and evolving perspective. We conclude by discussing growth opportunities and possible future directions of the RERC program.

Fall Detection in Individuals With Lower Limb Amputations Using Mobile Phones: Machine Learning Enhances Robustness for Real-World Applications

Background

Automatically detecting falls with mobile phones provides an opportunity for rapid response to injuries and better knowledge of what precipitated the fall and its consequences. This is beneficial for populations that are prone to falling, such as people with lower limb amputations. Prior studies have focused on fall detection in able-bodied individuals using data from a laboratory setting. Such approaches may provide a limited ability to detect falls in amputees and in real-world scenarios.

Objective

The aim was to develop a classifier that uses data from able-bodied individuals to detect falls in individuals with a lower limb amputation, while they freely carry the mobile phone in different locations and during free-living.

Methods

We obtained 861 simulated indoor and outdoor falls from 10 young control (non-amputee) individuals and 6 individuals with a lower limb amputation. In addition, we recorded a broad database of activities of daily living, including data from three participants’ free-living routines. Sensor readings (accelerometer and gyroscope) from a mobile phone were recorded as participants freely carried it in three common locations—on the waist, in a pocket, and in the hand. A set of 40 features were computed from the sensors data and four classifiers were trained and combined through stacking to detect falls. We compared the performance of two population-specific models, trained and tested on either able-bodied or amputee participants, with that of a model trained on able-bodied participants and tested on amputees. A simple threshold-based classifier was used to benchmark our machine-learning classifier.

Results

The accuracy of fall detection in amputees for a model trained on control individuals (sensitivity: mean 0.989, 1.96*standard error of the mean [SEM] 0.017; specificity: mean 0.968, SEM 0.025) was not statistically different (P=.69) from that of a model trained on the amputee population (sensitivity: mean 0.984, SEM 0.016; specificity: mean 0.965, SEM 0.022). Detection of falls in control individuals yielded similar results (sensitivity: mean 0.979, SEM 0.022; specificity: mean 0.991, SEM 0.012). A mean 2.2 (SD 1.7) false alarms per day were obtained when evaluating the model (vs mean 122.1, SD 166.1 based on thresholds) on data recorded as participants carried the phone during their daily routine for two or more days. Machine-learning classifiers outperformed the threshold-based one (P<.001).

Conclusions

A mobile phone-based fall detection model can use data from non-amputee individuals to detect falls in individuals walking with a prosthesis. We successfully detected falls when the mobile phone was carried across multiple locations and without a predetermined orientation. Furthermore, the number of false alarms yielded by the model over a longer period of time was reasonably low. This moves the application of mobile phone-based fall detection systems closer to a real-world use case scenario.

Activity Recognition in Individuals Walking With Assistive Devices: The Benefits of Device-Specific Models

BACKGROUND

Wearable sensors gather data that machine-learning models can convert into an identification of physical activities, a clinically relevant outcome measure. However, when individuals with disabilities upgrade to a new walking assistive device, their gait patterns can change, which could affect the accuracy of activity recognition.

OBJECTIVE

The objective of this study was to assess whether we need to train an activity recognition model with labeled data from activities performed with the new assistive device, rather than data from the original device or from healthy individuals.

METHODS

Data were collected from 11 healthy controls as well as from 11 age-matched individuals with disabilities who used a standard stance control knee-ankle-foot orthosis (KAFO), and then a computer-controlled adaptive KAFO (Ottobock C-Brace). All subjects performed a structured set of functional activities while wearing an accelerometer on their waist, and random forest classifiers were used as activity classification models. We examined both global models, which are trained on other subjects (healthy or disabled individuals), and personal models, which are trained and tested on the same subject.

RESULTS

Median accuracies of global and personal models trained with data from the new KAFO were significantly higher (61% and 76%, respectively) than those of models that use data from the original KAFO (55% and 66%, respectively) (Wilcoxon signed-rank test, P=.006 and P=.01). These models also massively outperformed a global model trained on healthy subjects, which only achieved a median accuracy of 53%. Device-specific models conferred a major advantage for activity recognition.

CONCLUSIONS

Our results suggest that when patients use a new assistive device, labeled data from activities performed with the specific device are needed for maximal precision activity recognition. Personal device-specific models yield the highest accuracy in such scenarios, whereas models trained on healthy individuals perform poorly and should not be used in patient populations.