Search

The main goal of this research is to develop a a machine learning based method in order to detect cough from acceleration signals. In this study, two different methods are proposed: a conventional one that uses XGBoost as a classifier and a deep learning which uses CNN-1D as an architecture. We found that these models were able to distinguish between acceleration signals caused by coughing and acceleration signals caused by other activities such as clearing throat, talking, laughing and movements in different directions with high accuracy. This study affirms that cough monitoring based on accelerometer measurements generated by the Hexoskin device is possible, making it a new user-independent tool for cough detection.

People with urinary incontinence (UI) often face a significant social stigma feeling ashamed of their condition and worrying about others discovering it. In order to improve the quality of life of those with incontinence, recent technological advancements enabled the development of non-invasive devices for detecting urinary leakage (UL). However, no comprehensive study has been conducted to state the most suitable types of sensors and the fundamental features necessary to design such devices, while also pointing gaps for future research. To address this, we conducted a mini review using four electronic databases limiting our search to English-written papers published in peer-reviewed journals. We retrieved articles that met the chosen inclusion criteria and classified them based on sensor type used, its location, the detection technique employed, and whether it was an e-Textile design and a reusable product or not. Across the studies, UL was detected using different approaches leading to heterogeneous results. Electrodes commonly used as sensing elements, along with textile as substrate material, and an indicator of UL based on resistance value, appeared to be widely exploited. However, the outcomes were not correlated with any specific type of UI. Consequently, we hypothesize that any non-invasive device could potentially be used for different types of UI. Nevertheless, further studies need to be conducted to confirm this statement. The designed literature mapping provides readers with an overview of the recent non-invasive wearable technologies in UL detection and offers a roadmap for future innovations.

PURPOSE: To identify and characterize phenotypes in healthy 3D knee kinematics. METHODS: Three-dimensional (3D) knee kinematics, namely varus/valgus, flexion/extension, and tibial external/internal rotation movements, were measured via knee kinesiography. A global approach based on a hierarchical discriminant model (BIRCH) was used to separate 3D knee kinematics into homogeneous clusters. Between-phenotype comparisons were made using intracluster correlation coefficient (ICC) and 1D-Statistical parametric mapping (for kinematics), ANOVA and chi-square tests (for participants’ characteristics). RESULTS: Three phenotypes were identified from 150 healthy knees (34 men, 41 women, 33 ± 10 years, Figure 1). First phenotype (blue curve) is characterized by predominant dynamic varus alignment, neutral axial rotation during stance, and more internal tibia rotation during initial swing. This group was composed mostly of men (60.4%). Phenotype 2 (orange), comprised mainly of women (70%), show predominant dynamic valgus alignment, an externally rotated tibia throughout gait cycle. Phenotype 3 (green) shows the most neutral frontal and transverse plane alignment, but greater flexion contracture at heel strike and throughout stance. Phenotype 2 had a higher body mass index than phenotype 1 (p < 0.05). Mean ICC within each phenotype was higher in frontal (0.67) and transverse (0.53) planes compared to sagittal (0.38), suggesting a greater contribution of these planes to identify distinct phenotypes. CONCLUSION: Healthy knee 3D kinematics may be divided in three phenotypes. This finding will help better understand the impact of knee disorders and therefore help define more personalized care plan.

Monitoring patients at risk of epileptic seizure is critical for optimal treatment and ensuing the reduction of seizure risk and complications. In general, seizure detection is done manually in hospitals and involves time-consuming visual inspection and interpretation by experts of electroencephalography (EEG) recordings. The purpose of this study is to investigate the pertinence of band-limited spectral power and signal complexity in order to discriminate between seizure and seizure-free EEG brain activity. The signal complexity and spectral power are evaluated in five frequency intervals, namely, the delta, theta, alpha, beta, and gamma bands, to be used as EEG signal feature representation. Classification of seizure and seizure-free data was performed by prevalent potent classifiers. Substantial comparative performance evaluation experiments were performed on a large EEG data record of 341 patients in the Temple University Hospital EEG seizure database. Based on statistically validated criteria, results show the efficiency of band-limited spectral power and signal complexity when using random forest and gradient-boosting decision tree classifiers (95% of the area under the curve (AUC) and 91% for both F-measure and accuracy). These results support the use of these automatic classification schemes to assist the practicing neurologist interpret EEG records more accurately and without tedious visual inspection.

One of the main motivations for classifying knee kinematic signals, which measure flexion/extension, abduction/adduction, and internal/external rotation angle variations during locomotion, is to assist diagnosis of knee joint pathologies. These signals are informative but of high dimensionality and high within-subject variability, serious difficulties which are often referred to as the curse of dimensionality. In general, current machine learning studies of knee pathology classification include feature extraction as an essential component, where feature design is time-consuming and generally not applicable to newly acquired data. To overcome such problems, the purpose of this study is to investigate classification of knee kinematic signals through the entire gait using deep neural networks. The signals are first pre-processed to identify representative patterns of a given subject using the within-subject variability evaluation for outliers’ removal and reliable curves’ selection. The patterns are then used for deep learning of discriminative classifiers. The proposed pre-processing method encompasses steps of gait events detection, normalization, outlier detection, and cycles’ selection. In order to measure the reliability of the subjects’ curves before and after pre-processing, we computed the intraclass correlation (ICC) estimates and their 95% confidence intervals for knee kinematics of a multicentric dataset of 239 subjects: 49 asymptomatic (AS) subjects and 190 knee osteoarthrosis (OA) patients. Then, we describe experiments which support the comparison of five deep neural networks, that have been performed on similar data, to distinguish between asymptomatic (AS) subjects and knee osteoarthrosis (OA) patients and justify further investigation in similar applications.

Measuring knee biomechanics provides valuable clinical information for defining patient-specific treatment options, including patient-oriented physical exercise programs. It can be done by a knee kinesiography test measuring the three-dimensional rotation angles (3D kinematics) during walking, thus providing objective knowledge about knee function in dynamic and weight-bearing conditions. The purpose of this study was to assess whether 3D kinematics can be efficiently used to predict the impact of a physical exercise program on the condition of knee osteoarthritis (OA) patients. The prediction was based on 3D knee kinematic data, namely flexion/extension, adduction/abduction and external/internal rotation angles collected during a treadmill walking session at baseline. These measurements are quantifiable information suitable to develop automatic and objective methods for personalized computer-aided treatment systems. The dataset included 221 patients who followed a personalized therapeutic physical exercise program for 6 months and were then assigned to one of two classes, Improved condition (I) and not-Improved condition (nI). A 10% improvement in pain was needed at the 6-month follow-up compared to baseline to be in the improved group. The developed model was able to predict I and nI with 84.4% accuracy for men and 75.5% for women using a decision tree classifier trained with 3D knee kinematic data taken at baseline and a 10-fold validation procedure. The models showed that men with an impaired control of their varus thrust and a higher pain level at baseline, and women with a greater amplitude of internal tibia rotation were more likely to report improvements in their pain level after 6 months of exercises. Results support the effectiveness of decision trees and the relevance of 3D kinematic data to objectively predict knee OA patients’ response to a treatment consisting of a physical exercise program.

The purpose of this study is to identify healthy phenotypes in knee kinematics based on clustering data analysis. Our analysis uses the 3D knee kinematics curves, namely, flexion/extension, abduction/adduction, and tibial internal/external rotation, measured via a KneeKG™ system during a gait task. We investigated two data representation approaches that are based on the joint analysis of the three dimensions. The first is a global approach that is considered a concatenation of the kinematic data without any dimensionality reduction. The second is a local approach that is considered a set of 69 biomechanical parameters of interest extracted from the 3D kinematic curves. The data representations are followed by a clustering process, based on the BIRCH (balanced iterative reducing and clustering using hierarchies) discriminant model, to separate 3D knee kinematics into homogeneous groups or clusters. Phenotypes were obtained by averaging those groups. We validated the clusters using inter-cluster correlation and statistical hypothesis tests. The simulation results showed that the global approach is more efficient, and it allows the identification of three descriptive 3D kinematic phenotypes within a healthy knee population.

Purpose: The diagnosis of knee osteoarthritis (OA) is typically well established with a clinical evaluation and confirmed with an X-Ray assessing the joint’ structural changes and disease progression. Guidelines also recommend taking into account mechanical factors (static and dynamic) to better understand knee function, since they may influence treatment outcomes. However, the relationship between clinical condition of the knee and biomechanical characteristics is not well known, including how such information stands compared to those from other conventional assessments, such as X-ray and physical assessment. The aim of this study is to evaluate the associations between the knee clinical condition assessed by patient-reported outcome measures and parameters from three different types of assessments, namely radiographic, musculoskeletal, and biomechanical assessment in OA patients. Methods: This cross-sectional study was conducted on patients with 1) knee pain ≥ 4/10 on a numeric rating scale in the past 7 days, 2) Kellgren-Lawrence (KL) radiographic OA severity grade higher than KL2, and 3) who were not on a waiting list for knee arthroplasty. Patients’ knee clinical condition was assessed using the Knee Injury and Osteoarthritis Outcome Score (KOOS) questionnaire that consists of five subscales: pain, symptoms, function in daily living (ADL), function in sport and recreation (Sport/Rec) and knee-related quality of life (QOL). Twenty musculoskeletal tests were performed by a therapist, including passive flexion and extension ranges of motion (ROM), muscle strength (10 tests assessing hip, knee, and ankle), flexibility (4 tests), swelling measured by the circumference difference between knees, effusion, balance, and functional 30-second chair stand tests (30s_CST). Finally, dynamic mechanical factors were measured during a knee kinesiography exam with the KneeKG™ system (Emovi Inc., QC, Canada) where 70 biomechanical parameters were extracted from 3D knee kinematic

Human activity recognition (HAR) by wearable sensor devices embedded in the Internet of things (IOT) can play a significant role in remote health monitoring and emergency notification to provide healthcare of higher standards. The purpose of this study is to investigate a human activity recognition method of accrued decision accuracy and speed of execution to be applicable in healthcare. This method classifies wearable sensor acceleration time series data of human movement using an efficient classifier combination of feature engineering-based and feature learning-based data representation. Leave-one-subject-out cross-validation of the method with data acquired from 44 subjects wearing a single waist-worn accelerometer on a smart textile, and engaged in a variety of 10 activities, yielded an average recognition rate of 90%, performing significantly better than individual classifiers. The method easily accommodates functional and computational parallelization to bring execution time significantly down.

Carrot (Daucus carota L.) is the most widely consumed root vegetable since it is an important source of nutritional compounds, mainly antioxidants and sugars. In Tunisia, despite the genetic diversity observed in carrot germplasm, including landraces and wild relatives, no research has been conducted on the biochemical composition of carrot. Thus, this study aims to analyse carotenoids, soluble sugars, total phenols, total flavonoids and colour properties of 14 carrot landraces, in order to determine the diversity among them and evaluate the relationships among their biochemical contents. The main carotenoids identified were a-carotene, B-carotene and lutein. Orange carrots were richer in B-carotene and a-carotene than yellow carrots. The major sugars were sucrose, glucose, fructose and galactose. Significant differences were observed among the Tunisian carrot landraces with respect to their biochemical composition and colour characteristics. Total carotenoids and total sugars ranged from 155.74 to 511.44 ug/g of dw and from 368.77 to 546.79 mg/g of dw, respectively. Total phenols and total flavonoids varied from 24.13 to 41.39 mg GAE/100 g of dw and from 16.51 to 24.85 ug CE/100 g of dw, respectively. Significant, positive and negative correlations were found among the measured parameters. A principal component analysis (PCA) and agglomerative hierarchical clustering (AHC) were performed to classify the Tunisian carrot landraces on the basis of colour properties and biochemical compounds. The PCA divided the landraces into four main groups and AHC classified them into two major clusters. The Tunisian carrot landraces were found to be rich in bioactive compounds; they could be good candidates for future breeding programs.

The analysis of knee kinematic data, which come in the form of a small sample of discrete curves that describe repeated measurements of the temporal variation of each of the knee three fundamental angles of rotation during a subject walking cycle, can inform knee pathology classification because, in general, different pathologies have different kinematic data patterns. However, high data dimensionality and the scarcity of reference data, which characterize this type of application, challenge classification and make it prone to error, a problem Duda and Hart refer to as the curse of dimensionality. The purpose of this study is to investigate a sample-based classifier which evaluates data proximity by the two-sample Hotelling T2 statistic. This classifier uses the whole sample of an individual’s measurements for a better support to classification, and the Hotelling T2 hypothesis testing made applicable by dimensionality reduction. This method was able to discriminate between femero-rotulian (FR) and femero-tibial (FT) knee osteoarthritis kinematic data with an accuracy of 88.1% , outperforming significantly current state-of-the-art methods which addressed similar problems. Extended to the much harder three-class problem involving pathology categories FR and FT, as well as category FR-FT which represents the incidence of both diseases FR and FT in a same individual, the scheme was able to reach a performance that justifies its further use and investigation in this and other similar applications.

Keywords: Knee Kinematic, Biomechanical Data, Feature Selection, Complexity Measures, Arthroplasty. Abstract: The purpose of this study is to investigate a method to select a set of knee kinematic data fatures to characterize surgical vs nonsurgical arthroplasty subjects. The kinematic features are generated from 3D knee kinematic data patterns, namely, rotations of flexion-extension, abduction-adduction, and tibial internal-external recorded during a walking task on a dedicated treadmill. The discrimination features are selected using three types of statistical complexity measures: the Fisher discriminant ratio, volume of overlap region, and feature efficiency. The interclass distance measurements which the features thus selected induce demonstrate their effectiveness to characterize surgical and nonsurgical subjects for arthroplasty.