Search

Abstract Background The research gap addressed in this study is the applicability of deep neural network (NN) models on wearable sensor data to recognize different activities performed by patients with Parkinson’s Disease (PwPD) and the generalizability of these models to PwPD using labeled healthy data. Methods The experiments were carried out utilizing three datasets containing wearable motion sensor readings on common activities of daily living. The collected readings were from two accelerometer sensors. PAMAP2 and MHEALTH are publicly available datasets collected from 10 and 9 healthy, young subjects, respectively. A private dataset of a similar nature collected from 14 PwPD patients was utilized as well. Deep NN models were implemented with varying levels of complexity to investigate the impact of data augmentation, manual axis reorientation, model complexity, and domain adaptation on activity recognition performance. Results A moderately complex model trained on the augmented PAMAP2 dataset and adapted to the Parkinson domain using domain adaptation achieved the best activity recognition performance with an accuracy of 73.02%, which was significantly higher than the accuracy of 63% reported in previous studies. The model’s F1 score of 49.79% significantly improved compared to the best cross-testing of 33.66% F1 score with only data augmentation and 2.88% F1 score without data augmentation or domain adaptation. Conclusion These findings suggest that deep NN models originating on healthy data have the potential to recognize activities performed by PwPD accurately and that data augmentation and domain adaptation can improve the generalizability of models in the healthy-to-PwPD transfer scenario. The simple/moderately complex architectures tested in this study could generalize better to the PwPD domain when trained on a healthy dataset compared to the most complex architectures used. The findings of this study could contribute to the development of accurate wearable-based activity monitoring solutions for PwPD, improving clinical decision-making and patient outcomes based on patient activity levels.

IntroductionAlzheimer's disease and related disorders (ADRD) progressively impair cognitive function, prompting the need for early detection to mitigate its impact. Mild Cognitive Impairment (MCI) may signal an early cognitive decline due to ADRD. Thus, developing an accessible, non-invasive method for detecting MCI is vital for initiating early interventions to prevent severe cognitive deterioration.MethodsThis study explores the utility of analyzing gait patterns, a fundamental aspect of human motor behavior, on straight and oval paths for diagnosing MCI. Using a Kinect v.2 camera, we recorded the movements of 25 body joints from 25 individuals with MCI and 30 healthy older adults (HC). Signal processing, descriptive statistical analysis, and machine learning techniques were employed to analyze the skeletal gait data in both walking conditions.Results and discussionThe study demonstrated that both straight and oval walking patterns provide valuable insights for MCI detection, with a notable increase in identifiable gait features in the more complex oval walking test. The Random Forest model excelled among various algorithms, achieving an 85.50% accuracy and an 83.9% F-score in detecting MCI during oval walking tests. This research introduces a cost-effective, Kinect-based method that integrates gait analysis—a key behavioral pattern—with machine learning, offering a practical tool for MCI screening in both clinical and home environments.

Accurate balance assessment is important in healthcare for identifying and managing conditions affecting stability and coordination. It plays a key role in preventing falls, understanding movement disorders, and designing appropriate therapeutic interventions across various age groups and medical conditions. However, traditional balance assessment methods often suffer from subjectivity, lack of comprehensive balance assessments and remote assessment capabilities, and reliance on specialized equipment and expert analysis. In response to these challenges, our study introduces an innovative approach for estimating scores on the Modified Clinical Test of Sensory Interaction on Balance (m-CTSIB). Utilizing wearable sensors and advanced machine learning algorithms, we offer an objective, accessible, and efficient method for balance assessment. We collected comprehensive movement data from 34 participants under four different sensory conditions using an array of inertial measurement unit (IMU) sensors coupled with a specialized system to evaluate ground truth m-CTSIB balance scores for our analysis. This data was then preprocessed, and an extensive array of features was extracted for analysis. To estimate the m-CTSIB scores, we applied Multiple Linear Regression (MLR), Support Vector Regression (SVR), and XGBOOST algorithms. Our subject-wise Leave-One-Out and 5-Fold cross-validation analysis demonstrated high accuracy and a strong correlation with ground truth balance scores, validating the effectiveness and reliability of our approach. Key insights were gained regarding the significance of specific movements, feature selection, and sensor placement in balance estimation. Notably, the XGBOOST model, utilizing the lumbar sensor data, achieved outstanding results in both methods, with Leave-One-Out cross-validation showing a correlation of 0.96 and a Mean Absolute Error (MAE) of 0.23 and 5-fold cross-validation showing comparable results with a correlation of 0.92 and an MAE of 0.23, confirming the model’s consistent performance. This finding underlines the potential of our method to revolutionize balance assessment practices, particularly in settings where traditional methods are impractical or inaccessible.

The Unified Parkinson’s Disease Rating Scale (UPDRS) is used to recognize patients with Parkinson’s disease (PD) and rate its severity. The rating is crucial for disease progression monitoring and treatment adjustment. This study aims to advance the capabilities of PD management by developing an innovative framework that integrates deep learning with wearable sensor technology to enhance the precision of UPDRS assessments. We introduce a series of deep learning models to estimate UPDRS Part III scores, utilizing motion data from wearable sensors. Our approach leverages a novel Multi-shared-task Self-supervised Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) framework that processes raw gyroscope signals and their spectrogram representations. This technique aims to refine the estimation accuracy of PD severity during naturalistic human activities. Utilizing 526 min of data from 24 PD patients engaged in everyday activities, our methodology demonstrates a strong correlation of 0.89 between estimated and clinically assessed UPDRS-III scores. This model outperforms the benchmark set by single and multichannel CNN, LSTM, and CNN-LSTM models and establishes a new standard in UPDRS-III score estimation for free-body movements compared to recent state-of-the-art methods. These results signify a substantial step forward in bioengineering applications for PD monitoring, providing a robust framework for reliable and continuous assessment of PD symptoms in daily living settings.

The analysis of functional near-infrared spectroscopy (fNIRS) signals has not kept pace with the increased use of fNIRS in the behavioral and brain sciences. The popular grand averaging method collapses the oxygenated hemoglobin data within a predefined time of interest window and across multiple channels within a region of interest, potentially leading to a loss of important temporal and spatial information. On the other hand, the tensor decomposition method can reveal patterns in the data without making prior assumptions of the hemodynamic response and without losing temporal and spatial information. The aim of the current study was to examine whether the tensor decomposition method could identify significant effects and novel patterns compared to the commonly used grand averaging method for fNIRS signal analysis. We used two infant fNIRS datasets and applied tensor decomposition (i.e., canonical polyadic and Tucker decompositions) to analyze the significant differences in the hemodynamic response patterns across conditions. The codes are publicly available on GitHub. Bayesian analyses were performed to understand interaction effects. The results from the tensor decomposition method replicated the findings from the grand averaging method and uncovered additional patterns not detected by the grand averaging method. Our findings demonstrate that tensor decomposition is a feasible alternative method for analyzing fNIRS signals, offering a more comprehensive understanding of the data and its underlying patterns.

Alzheimer’s disease (AD) is a progressive neurodegenerative disease affecting cognitive and functional abilities. However, many patients presume lower cognitive or functional abilities because of aging and do not undergo clinical assessments until the symptoms become too advanced. Developing a low-cost and easy-to-use AD detection tool, which can be used in any clinical or non-clinical setting, can enable widespread AD assessments and diagnosis. This paper investigated the feasibility of developing such a tool to detect AD vs. healthy control (HC) from a simple balance and walking assessment called the Timed Up and Go (TUG) test. We collected joint position data of 47 HC and 38 AD subjects as they performed TUG in front of a Kinect V.2 camera. Our signal processing and statistical analyses provided a comprehensive analysis of balance and gait with 12 significant features for discriminating AD from HC after adjusting for age and the Geriatric Depression Scale. Using these features and a support vector machine classifier, our model classified the two groups with an average accuracy of 97.75% and an F-score of 97.67% for five-fold cross-validation and 98.68% and 98.67% for leave-one-subject out cross-validation. These results demonstrate the potential of our approach as a new quantitative complementary tool for detecting AD among older adults. Our work is novel as it presents the first application of Kinect V.2 camera and machine learning to provide a comprehensive and quantitative analysis of the TUG test to detect AD patients from HC. This study supports the feasibility of developing a low-cost and convenient AD assessment tool that can be used during routine checkups or even at home; however, future investigations could confirm its clinical diagnostic value in a larger cohort.

Abstract The early detection of the coronavirus disease 2019 (COVID-19) outbreak is important to save people’s lives and restart the economy quickly and safely. People’s social behavior, reflected in their mobility data, plays a major role in spreading the disease. Therefore, we used the daily mobility data aggregated at the county level beside COVID-19 statistics and demographic information for short-term forecasting of COVID-19 outbreaks in the United States. The daily data are fed to a deep learning model based on Long Short-Term Memory (LSTM) to predict the accumulated number of COVID-19 cases in the next two weeks. A significant average correlation was achieved (r=0.83 (p = 0.005)) between the model predicted and actual accumulated cases in the interval from August 1, 2020 until January 22, 2021. The model predictions had r > 0.7 for 87% of the counties across the United States. A lower correlation was reported for the counties with total cases of

Abstract Levodopa-induced dyskinesias are abnormal involuntary movements experienced by the majority of persons with Parkinson’s disease (PwP) at some point over the course of the disease. Choreiform as the most common phenomenology of levodopa-induced dyskinesias can be managed by adjusting the dose/frequency of PD medication(s) based on a PwP’s motor fluctuations over a typical day. We developed a sensor-based assessment system to provide such information. We used movement data collected from the upper and lower extremities of 15 PwPs along with a deep recurrent model to estimate dyskinesia severity as they perform different activities of daily living (ADL). Subjects performed a variety of ADLs during a 4-h period while their dyskinesia severity was rated by the movement disorder experts. The estimated dyskinesia severity scores from our model correlated highly with the expert-rated scores (r = 0.87 (p < 0.001)), which was higher than the performance of linear regression that is commonly used for dyskinesia estimation (r = 0.81 (p < 0.001)). Our model provided consistent performance at different ADLs with minimum r = 0.70 (during walking) to maximum r = 0.84 (drinking) in comparison to linear regression with r = 0.00 (walking) to r = 0.76 (cutting food). These findings suggest that when our model is applied to at-home sensor data, it can provide an accurate picture of changes of dyskinesia severity facilitating effective medication adjustments.

Abstract Background Unified Parkinson Disease Rating Scale-part III (UPDRS III) is part of the standard clinical examination performed to track the severity of Parkinson’s disease (PD) motor complications. Wearable technologies could be used to reduce the need for on-site clinical examinations of people with Parkinson’s disease (PwP) and provide a reliable and continuous estimation of the severity of PD at home. The reported estimation can be used to successfully adjust the dose and interval of PD medications. Methods We developed a novel algorithm for unobtrusive and continuous UPDRS-III estimation at home using two wearable inertial sensors mounted on the wrist and ankle. We used the ensemble of three deep-learning models to detect UPDRS-III-related patterns from a combination of hand-crafted features, raw temporal signals, and their time–frequency representation. Specifically, we used a dual-channel, Long Short-Term Memory (LSTM) for hand-crafted features, 1D Convolutional Neural Network (CNN)-LSTM for raw signals, and 2D CNN-LSTM for time–frequency data. We utilized transfer learning from activity recognition data and proposed a two-stage training for the CNN-LSTM networks to cope with the limited amount of data. Results The algorithm was evaluated on gyroscope data from 24 PwP as they performed different daily living activities. The estimated UPDRS-III scores had a correlation of $$0.79\, (\textit{p}

Automated momentary estimation of positive and negative affects (PA and NA), the basic sense of feeling, can play an essential role in detecting the early signs of mood disorders. Physiological wearable sensors and machine learning have a potential to make such automated and continuous measurements. However, the physiological signals' features that are associated with the subject-reported PA or NA may not be known. In this work, we use data-driven feature extraction based on deep learning to investigate the application of raw physiological signals for estimating PA and NA. Specifically, we propose two multi-modal data fusion methods with deep Convolutional Neural Networks. We use the proposed architecture to estimate PA and NA and also classify baseline, stress, and amusement emotions. The training and evaluation of the methods are performed using four physiological and one chest motion signal modalities collected using a chest sensing unit from 15 subjects. Overall, our proposed model performed better than traditional machine learning on hand-crafted features. Utilizing only two modalities, our proposed model estimated PA with a correlation of 0.69 (P

Abstract Background Catheter ablation therapy involving isolation of pulmonary veins (PVs) from the left atrium is performed to terminate atrial fibrillation (AF). Unfortunately, standalone PV isolation procedure has shown to be a suboptimal success with AF continuation or recurrence. One reason, especially in patients with persistent or high-burden paroxysmal AF, is known to be due to the formation of repeating-pattern AF sources with a meandering core inside the atria. However, there is a need for accurate mapping and localization of these sources during catheter ablation. Methods A novel AF source area probability (ASAP) mapping algorithm was developed and evaluated in 2D and 3D atrial simulated tissues with various arrhythmia scenarios and a retrospective study with three cases of clinical human AF. The ASAP mapping analyzes the electrograms collected from a multipole diagnostic catheter that is commonly used during catheter ablation procedure to intelligently sample the atria and delineate the trajectory path of a meandering repeating-pattern AF source. ASAP starts by placing the diagnostic catheter at an arbitrary location in the atria. It analyzes the recorded bipolar electrograms to build an ASAP map over the atrium anatomy and suggests an optimal location for the subsequent catheter location. ASAP then determines from the constructed ASAP map if an AF source has been delineated. If so, the catheter navigation is stopped and the algorithm provides the area of the AF source. Otherwise, the catheter is navigated to the suggested location, and the process is continued until an AF-source area is delineated. Results ASAP delineated the AF source in over 95% of the simulated human AF cases within less than eight catheter placements regardless of the initial catheter placement. The success of ASAP in the clinical AF was confirmed by the ablation outcomes and the electrogram patterns at the delineated area. Conclusion Our analysis indicates the potential of the ASAP mapping to provide accurate information about the area of the meandering repeating-pattern AF sources as AF ablation targets for effective AF termination. Our algorithm could improve the success of AF catheter ablation therapy by locating and subsequently targeting patient-specific and repeating-pattern AF sources inside the atria.

Nowadays, portable and wireless wearable sensors have been commonly incorporated into the signal acquisition modules of healthcare monitoring systems [...]

BackgroundT wave alternans (TWA) is an electrocardiographic marker of heightened sudden death risk from ventricular tachyarrhythmias in patients with cardiomyopathy. TWA is evaluated from the 12-lead electrocardiogram, Frank lead, or Holter lead recordings, however these clinical lead configurations will not record TWA from adjacent regions of the body torso.ObjectiveWe tested the hypothesis that changing heart rate or ventricular activation may alter the body surface distribution of TWA such that the clinical ECG leads fail to detect TWA in some patients; thereby producing a false-negative test.MethodsIn 28 cardiomyopathy patients (left ventricular ejection fraction 28±6%), 114 unipolar electrograms were recorded across the body torso during incremental atrial pacing, followed by atrioventricular pacing at 100, 110 and 120bpm. TWA was measured from each unipolar electrogram using the spectral method. A clinically positive TWA test was defined as TWA magnitude (Valt) ≥1.9 uV with k ≥3 at ≤110bpm.ResultsMaximum Valt (TWAmax) was greater from the body torso than clinical leads during atrial (pConclusionsThe body surface distribution of TWA is modulated by heart rate and the sequence of ventricular activation in patients with cardiomyopathy, which can give rise to modest false-negative TWA signal detection using standard clinical leads.

Traumatic brain injuries may cause intracranial hemorrhages (ICH). ICH could lead to disability or death if it is not accurately diagnosed and treated in a time-sensitive procedure. The current clinical protocol to diagnose ICH is examining Computerized Tomography (CT) scans by radiologists to detect ICH and localize its regions. However, this process relies heavily on the availability of an experienced radiologist. In this paper, we designed a study protocol to collect a dataset of 82 CT scans of subjects with a traumatic brain injury. Next, the ICH regions were manually delineated in each slice by a consensus decision of two radiologists. The dataset is publicly available online at the PhysioNet repository for future analysis and comparisons. In addition to publishing the dataset, which is the main purpose of this manuscript, we implemented a deep Fully Convolutional Networks (FCNs), known as U-Net, to segment the ICH regions from the CT scans in a fully-automated manner. The method as a proof of concept achieved a Dice coefficient of 0.31 for the ICH segmentation based on 5-fold cross-validation.

Tremor is one of the main symptoms of Parkinson’s Disease (PD) that reduces the quality of life. Tremor is measured as part of the Unified Parkinson Disease Rating Scale (UPDRS) part III. However, the assessment is based on onsite physical examinations and does not fully represent the patients’ tremor experience in their day-to-day life. Our objective in this paper was to develop algorithms that, combined with wearable sensors, can estimate total Parkinsonian tremor as the patients performed a variety of free body movements. We developed two methods: an ensemble model based on gradient tree boosting and a deep learning model based on long short-term memory (LSTM) networks. The developed methods were assessed on gyroscope sensor data from 24 PD subjects. Our analysis demonstrated that the method based on gradient tree boosting provided a high correlation (r = 0.96 using held-out testing and r = 0.93 using subject-based, leave-one-out cross-validation) between the estimated and clinically assessed tremor subscores in comparison to the LSTM-based method with a moderate correlation (r = 0.84 using held-out testing and r = 0.77 using subject-based, leave-one-out cross-validation). These results indicate that our approach holds great promise in providing a full spectrum of the patients’ tremor from continuous monitoring of the subjects’ movement in their natural environment.

The success of medication adjustment in Parkinson’s disease (PD) patients with motor fluctuation relies on the knowledge about their fluctuation severity. However, because of the temporal and spatial variability in motor fluctuations, a single clinical examination often fails to capture the spectrum of motor impairment experienced in routine daily life. In this study, we developed an algorithm to estimate the degree of motor fluctuation severity from two wearable sensors’ data during subjects’ free body movements. Specifically, we developed a new hybrid feature extraction method to represent the longitudinal changes of motor function from the sensor data. Next, we developed a classification model based on random forest to learn the changes in the patterns of the sensor data as the severity of the motor function changes. We evaluated our algorithm using data from 24 subjects with idiopathic PD as they performed a variety of daily routine activities. A leave-one-subject-out assessment of the algorithm resulted in 83.33% accuracy, indicating that our approach holds a great promise to passively detect degree of motor fluctuation severity from continuous monitoring of an individual’s free body movements. Such a sensor-based assessment system and algorithm combination could provide the objective and comprehensive information about the fluctuation severity that can be used by the treating physician to effectively adjust therapy for PD patients with troublesome motor fluctuation.

The number of people affected by speech problems is increasing as the modern world places increasing demands on the human voice via mobile telephones, voice recognition software, and interpersonal verbal communications. In this paper, we propose a novel methodology for automatic pattern classification of pathological voices. The main contribution of this paper is extraction of meaningful and unique features using Adaptive time-frequency distribution (TFD) and nonnegative matrix factorization (NMF). We construct Adaptive TFD as an effective signal analysis domain to dynamically track the nonstationarity in the speech and utilize NMF as a matrix decomposition (MD) technique to quantify the constructed TFD. The proposed method extracts meaningful and unique features from the joint TFD of the speech, and automatically identifies and measures the abnormality of the signal. Depending on the abnormality measure of each signal, we classify the signal into normal or pathological. The proposed method is applied on the Massachusetts Eye and Ear Infirmary (MEEI) voice disorders database which consists of 161 pathological and 51 normal speakers, and an overall classification accuracy of 98.6% was achieved.