Search

BackgroundUnified 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.MethodsWe 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.ResultsThe 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}0.79(p<0.0001)and a mean absolute error of 5.95 with the clinical examination scores without requiring the patients to perform any specific tasks.ConclusionOur analysis demonstrates the potential of our algorithm for estimating PD severity scores unobtrusively at home. Such an algorithm could provide the required motor-complication measurements without unnecessary clinical visits and help the treating physician provide effective management of the disease.

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

Introduction Targeting repeating‐pattern atrial fibrillation (AF) sources (reentry or focal drivers) can help in patient‐specific ablation therapy for AF; however, the development of reliable and accurate tools for locating such sources remains a major challenge. We describe iterative catheter navigation (ICAN) algorithm to locate AF drivers using a conventional circular Lasso catheter. Methods and Results At each step, the algorithm analyzes 10 bipolar electrograms recoded at a given catheter location and the history of previous catheter movements to determine if the source is inside the catheter loop. If not, it calculates new coordinates and selects a new position for the catheter. The process continues until a source is located. The algorithm was evaluated in a computer model of atrial tissue with various degrees of fibrosis under a broad range of arrhythmia scenarios. The latter included slow and fast reentry, macroreentry, figure‐of‐eight reentry, and fibrillatory conduction. Depending on the initial distance of the catheter from the source and scenario, it took about 3 to 16 steps to localize an AF source. In 94% of cases, the identified location was within 4 mm from the source, independently of the initial position of the catheter. The algorithm worked equally well in the presence of patchy fibrosis, low‐voltage areas, fragmented electrograms, and dominant‐frequency gradients. Conclusions AF repeating‐pattern sources can be localized using circular catheters without the need to map the entire tissue. The proposed algorithm has the potential to become a useful tool for patient‐specific ablation of AF sources located outside the pulmonary veins.

Objective Early detection of mild cognitive impairment (MCI) and Alzheimer’s disease (AD) can increase access to treatment and assist in advance care planning. However, the development of a diagnostic system that d7oes not heavily depend on cognitive testing is a major challenge. We describe a diagnostic algorithm based solely on gait and machine learning to detect MCI and AD from healthy. Methods We collected “single-tasking” gait (walking) and “dual-tasking” gait (walking with cognitive tasks) from 32 healthy, 26 MCI, and 20 AD participants using a computerized walkway. Each participant was assessed with the Montreal Cognitive Assessment (MoCA). A set of gait features (e.g., mean, variance and asymmetry) were extracted. Significant features for three classifications of MCI/healthy, AD/healthy, and AD/MCI were identified. A support vector machine model in a one-vs.-one manner was trained for each classification, and the majority vote of the three models was assigned as healthy, MCI, or AD. Results The average classification accuracy of 5-fold cross-validation using only the gait features was 78% (77% F1-score), which was plausible when compared with the MoCA score with 83% accuracy (84% F1-score). The performance of healthy vs. MCI or AD was 86% (88% F1-score), which was comparable to 88% accuracy (90% F1-score) with MoCA. Conclusion Our results indicate the potential of machine learning and gait assessments as objective cognitive screening and diagnostic tools. Significance Gait-based cognitive screening can be easily adapted into clinical settings and may lead to early identification of cognitive impairment, so that early intervention strategies can be initiated.

Functional near-infrared spectroscopy (fNIRS) provides an effective tool in neuroscience studies of cognition in infants. fNIRS signals are normally processed by applying ANOVA analysis on the grand average of the hemodynamic responses to investigate the cognitive-related differences between experimental groups. However, this averaging approach does not account for any differences in the temporal patterns of the responses. Therefore, we propose a new approach based on a combination of tensor decomposition and ANOVA. First, a four-way tensor of the hemodynamic responses is arranged as time × frequency × channel× subject and decomposed using Canonical Polyadic Decomposition (CPD). Next, ANOVA is applied to identify significant patterns between subjects. Instead of averaging, the CPD can capture the distinct patterns between groups in all the dimensions. We used fNIRS dataset of 70 infants who participated in an experiment to investigate cortical activation to an agent (i.e., mechanical claws vs. human hands) with different events (i.e., function and non-function). In the comparison with the traditional ANOVA, CPD+ANOVA identified the same significance factors. However, CPD+ANOVA discovered new information on the temporal and spatial patterns indicating a longer interval hemodynamic responses, which was missed using the traditional ANOVA. This new analysis of hemodynamic responses as captured using fNIRS will improve neuroscience and cognitive studies.

Alzheimer’s disease (AD) affects approximately 30 million people worldwide, and this number is predicted to triple by 2050 unless further discoveries facilitate the early detection and prevention of the disease. Computerized walkways for simultaneous assessment of motor-cognitive performance, known as a dual-task assessment, has been used to associate changes in gait characteristics to mild cognitive impairment (MCI) with early-stage disease. However, to our best knowledge, there is no validated method to detect MCI using the collective analysis of these gait characteristics. In this paper, we develop a machine learning approach to analyze the gait data from the dual-task assessment in order to detect subjects with cognitive impairment from healthy individuals. We collected dual-task gait data from a computerized walkway of a total of 92 subjects with 31 healthy control (HC) and 61 MCI. Using support vector machine (SVM) and gradient tree boosting, we developed a classifier to differentiate MCI from HC subjects and compared the results with a paper-based questionnaire assessment that has been commonly used in clinical practice. SVM provided the highest accuracy of 77.17% with 81.97% sensitivity and 67.74% specificity. Our results indicate the potential of machine learning + dual-task assessment to enable early diagnosis of cognitive decline before it advances to dementia and AD, so that early intervention or prevention strategies can be initiated.

Visual examination forms an integral part of cervical cancer screening. With the recent rise in smartphone-based health technologies, capturing cervical images using a smartphone camera for telemedicine and automated screening is gaining popularity. However, such images are highly prone to image corruption, typically out-of-focus target or camera shake blur. In this paper, we applied a generative adversarial network (GAN) to deblur mobile-phone cervical (MC) images, and we evaluate the deblur quality using various measures. Our evaluation process is three-fold: first, we calculate the peak signal to noise ratio (PSNR) and the structural similarity (SSIM) of a test dataset with ground truth availability. Next, we calculate the perception based image quality evaluator (PIQE) score of a test dataset without ground truth availability. Finally, we classify a dataset of blurred and the corresponding deblurred images into normal/abnormal MC images. The resulting change in classification accuracy was our final assessment. Our evaluation experiments show that deblurring of MC images can potentially improve the accuracy of both manual and automated cancerous lesion screening.

Image augmentation is a commonly performed technique to prevent class imbalance in datasets to compensate for insufficient training samples, or to prevent model overfitting. Traditional augmentation (TA) techniques include various image transformations, such as rotation, translation, channel splitting, etc. Alternatively, Generative Adversarial Network (GAN), due to its proven ability to synthesize convincingly-realistic images, has been used to perform image augmentation as well. However, it is unclear whether GAN augmentation (GA) strategy provides an advantage over TA for medical image classification tasks. In this paper, we study the usefulness of TA and GA for classifying abnormal chest X-ray (CXR) images. We first trained a progressive-growing GAN (PG-GAN) to synthesize high-resolution CXRs for performing GA. Then, we trained an abnormality classifier using three training sets individually - training set with TA, with GA and with no augmentation (NA). Finally, we analyzed the abnormality classifier's performance for the three training cases, which led to the following conclusions: (1) GAN strategy is not always superior to TA for improving the classifier's performance; (2) in comparison to NA, however, both TA and GA leads to a significant performance improvement; and, (3) increasing the quantity of images in TA and GA strategies also improves the classifier's performance.

Motor fluctuations between “OFF” state (with no benefit from medication) and “ ON” state (with optimum benefit from medication) are a major focus of clinical managements in individuals with mid-stage and advance Parkinson’s disease (PD). In this work, an automated algorithm based on Long Short-Term Memory (LSTM) as a deep learning method is developed to identify motor fluctuations in individuals with PD using wearable sensors during a variety of daily living activities. This network was evaluated on two datasets i.e., Dataset 1 and Dataset 2) that included recordings of 19 individuals with PD using subject-based leave-one-out cross-validation. The designed LSTM network yielded promising results using only one ankle sensor with an average classification rate of 73% and 77% for Dataset 1 and Dataset 2, respectively.

Catheter ablation therapy involving isolation of pulmonary veins (PVs) remains the cornerstone procedure to treat AF. However, due to the sub-optimal success rates of PV isolation, there is a need for new ablation techniques to locate AF ablation targets known as rotors, outside of the PVs. In this paper, we developed a novel rotor-mapping algorithm that uses a conventional diagnostic catheter, Lasso, to locate a rotor source. The algorithm, called the Region of Rotor (ROR) Mapping, utilizes the characteristics of local bipolar electrograms to navigate the catheter's iterative placements while generating a map, overlaid on the atrial anatomy, that displays the potential rotor region. We evaluated the developed ROR mapping algorithm using a 2D simulation of AF on a tissue with heterogeneous conduction properties. The results demonstrated a significant success rate of 93% in accurately locating the region of the rotor with a mean distance of 1.4mm from the ground truth trajectory. The algorithm could play a critical role in mapping non-PV AF ablation targets and improving the outcome of AF ablation.

Motor fluctuations are a major focus of clinical managements in patients with mid-stage and advance Parkinson's disease (PD). In this paper, we develop a new patient-specific algorithm that can classify those fluctuations during a variety of activities. We extract a set of temporal and spectral features from the ambulatory signals and then introduce a semi-supervised classification algorithm based on K-means and self-organizing tree map clustering methods. Two different types of cluster labeling are introduced: hard and fuzzy labeling. The developed algorithm is evaluated on a dataset from triaxial gyroscope sensors for 12 PD patients. The average result of using K-means and fuzzy labeling on the trunk and the more affected leg sensors' readings was 75.96%, 70.57%, and 86.93% for accuracy, sensitivity, and specificity, respectively. The accuracy for individual patients varied from 99.95% to 42.53%, which was correlated with dyskinesia severity and the improvement of the PD symptoms with medication.

Parkinson disease (PD) cure remains one of the greatest challenges in chronic neurological disorder therapy, motivating efforts to provide actionable information to guide self-managed therapy adjustments. In this paper, we develop a data fusion approach to combine multi-dimensional data from body-worn inertial sensors to automatically identify different medication states of patients with PD. The proposed approach employs several signal processing algorithms including time-frequency analysis and tensor decomposition to extract features that can represent spectral, temporal, and spatial behavior of body motion data. A Support Vector Machines (SVM) classifier is trained and tested using a data set of 12 patients, which resulted in an overall average accuracy of 78%.

Atrial fibrillation (AF) is an irregular heart rhythm that can lead to stroke and other heart-related complications. Catheter ablation has been commonly used to destroy triggering sources of AF in the atria and consequently terminate the arrhythmia. However, efficient and accurate localization of the AF sustaining sources known as rotors is a major challenge in catheter ablation. In this paper, we developed a novel probabilistic algorithm that can adaptively guide a Lasso diagnostic catheter to locate the center of a rotor. Our algorithm uses a Bayesian updating approach to search for and locate rotors based on the characteristics of electrogram signals collected at every catheter placement. The algorithm was evaluated using a 10 × 10 cm 2 D atrial tissue simulation of the Nygren human atrial cell model and was able to successfully guide the catheter to the rotor center in 3.37 ± 1.05 (mean±std) steps (including placement at the center) when starting from any location on the tissue. Our novel automated algorithm can potentially play a significant role in patient-specific ablation of AF sources and increase the success of AF elimination procedures.

Time-frequency (TF) representations have been widely used over the past decade to characterize the non-stationary content of signals in the joint time and frequency domain. Although a number of effective TF analysis methods based on wavelet or Gabor transform have been developed, these methods use pre-determined basis functions and still require feature extraction methods to reduce redundancy and preserve important TF information related to the application of interest. This paper explores a novel TF feature extraction algorithm using a modified dictionary learning approach. The proposed algorithm is developed to modify learned dictionaries and derive TF features unique to each class. It emulates the way joint dictionary learning algorithms use common dictionaries to promote discrimination between the data from different classes, thereby allowing for an improved analysis of complex and multitask data. The proposed method indicated a significant performance in identification of the discriminant vs. common structures of the TF data.

Audio feature extraction and classification are important tools for audio signal analysis in many applications, such as multimedia indexing and retrieval, and auditory scene analysis. However, due to the nonstationarities and discontinuities exist in these signals, their quantification and classification remains a formidable challenge. In this paper, we develop a new approach for audio feature extraction to effectively quantify these nonstationarities in an attempt to achieve high classification accuracy for environmental audio signals. Our approach consists of three stages: first we propose to construct the time-frequency matrix (TFM) of audio signals using matching-pursuit time-frequency distribution (MP-TFD) technique, and then apply the non-negative matrix decomposition (NMF) technique to decompose the TFM into its significant components. Finally, we propose seven novel features from the spectral and temporal structures of the decomposed vectors in a way that they successfully represent joint TF structure of the audio signal, and combine them with the Mel-frequency cepstral coefficients (MFCCs) features. These features are examined using a database of 192 environmental audio signals which includes 20 aircraft, 17 helicopter, 20 drum, 15 flute, 20 piano, 20 animal, 20 bird, and 20 insect sounds, and the speech of 20 males and 20 females. The results of the numerical simulation support the effectiveness of the proposed approach for environmental audio classification with over 10% accuracy-rate improvement compared to the MFCC features.

Current approaches to improve the pattern recognition performance mainly focus on either extracting non-stationary and discriminant features of each class, or employing complex and nonlinear feature classifiers. However, little attention has been paid to the integration of these two approaches. Combining non-stationary feature analysis with complex feature classifiers, this article presents a novel direction to enhance the discriminatory power of pattern recognition methods. This approach, which is based on a fusion of non-stationary feature analysis with clustering techniques, proposes an algorithm to adaptively identify the feature vectors according to their importance in representing the patterns of discrimination. Non-stationary feature vectors are extracted using a non-stationary method based on time–frequency distribution and non-negative matrix factorization. The clustering algorithms including the K-means and self-organizing tree maps are utilized as unsupervised clustering methods followed by a supervised labeling. Two labeling methods are introduced: hard and fuzzy labeling. The article covers in detail the formulation of the proposed discriminant feature clustering method. Experiments performed with pathological speech classification, T-wave alternans evaluation from the surface electrocardiogram, audio scene analysis, and telemonitoring of Parkinson’s disease problems produced desirable results. The outcome demonstrates the benefits of non-stationary feature fusion with clustering methods for complex data analysis where existing approaches do not exhibit a high performance.

Automated classification of normal and pathological speech signals can provide an objective and accurate mechanism for pathological speech diagnosis, and is an active area of research. A large part of this research is based on analysis of acoustic measures extracted from sustained vowels. However, sustained vowels do not reflect real-world attributes of voice as effectively as continuous speech, which can take into account important attributes of speech such as rapid voice onset and termination, changes in voice frequency and amplitude, and sudden discontinuities in speech. This paper presents a methodology based on empirical mode decomposition (EMD) for classification of continuous normal and pathological speech signals obtained from a well-known database. EMD is used to decompose randomly chosen portions of speech signals into intrinsic mode functions, which are then analyzed to extract meaningful temporal and spectral features, including true instantaneous features which can capture discriminative information in signals hidden at local time-scales. A total of six features are extracted, and a linear classifier is used with the feature vector to classify continuous speech portions obtained from a database consisting of 51 normal and 161 pathological speakers. A classification accuracy of 95.7 % is obtained, thus demonstrating the effectiveness of the methodology.

The objective of this paper is to propose pole modeling-based features for audio classification in order to achieve a high classification performance. This paper investigates the suitable pole modeling computation method and evaluates the proposed audio features in an audio database with 40 human speech samples, and 40 non human audio signals including aircraft, helicopter, drum, flutes, and piano sounds. An accuracy rate of 85% is acheived using the pole modeling features and linear discriminant analysis (LDA). We also compare the performance of the pole modeling features with two well-known audio features: Autoregressive (AR), and Mel-frequency Cepstral coefficients (MFCCs). We found that pole modeling is an appropriate tool for real-time audio scene analysis.

Time-frequency matrix (TFM) decomposition using non-negative matrix factorization (NMF) has been recently considered as a successful tool for time-frequency (TF) quantification. In this paper, we modify the constraints of traditional cost function of NMF to make the method a better fit for TF quantification, and denote the new method with NMF discriminant base (NMFDB) decomposition. We evaluate the proposed method, and show that it successfully identifies the discriminant bases. Additionally, we measure the discrimination ability of NMFDB over the signals with very low discriminations, and compare it with the discrimination of the decomposed bases derived using traditional NMF. It is concluded that the proposed method is able to locate the region of difference with 20% better performance compared to the conventional NMF.

T wave alternans (TWA) has been associated with ventricular arrhythmias. Hence, TWA detection can risk stratify patients with heart disease who may experience sudden death from ventricular arhythmias. However, accurate TWA detection is technically challenging due to the low microvolt TWA signal and the confounding effect of biological noise such as movement, myopotentials or respiration. In this paper, we propose nonnegative matrix factorization (NMF)-Adaptive spectral method to increase the robustness of TWA detection in ambulatory electrocardiograms (ECGs). The proposed method applies a non-linear time-frequency (TF) analysis and NMF to the aligned ST-T waveforms. This method separates the TWA signal from the other non-desired ECG signal components, and detects TWA with high accuracy. The performance of our proposed method is validated in a clinical study using ECGs which confirms a TWA detection of 92% compared to 47% using the conventional spectral method.

Many feature extraction techniques in literature have studied data representation, but most techniques do not explicitly investigate the feature localization aspect. This is one of the first works in which signals have been transformed to matrices using positive time-frequency transform, and matrix decomposition and representation techniques such as PCA, ICA, and NMF have been applied on these matrices to study the feature representation and localization issues. To estimate each techniquespsila localization, we propose a localization measurement method. We also construct a non stationary synthetic signal which resembles major characteristics of real world signals, and then apply the feature extraction techniques on a simple time-frequency distribution (TFD) of this signal. The localization results show that matrix factorization 1-D deconvolution (NMF1D) offers the most localized features with 99.6% localization. In addition, we demonstrate that under different number of basis components and noisy conditions, NMF1D offers the most robust localization.

In this book chapter, we present an overview of our time-frequency (TF) based audio watermarking methods. First, a motivation on the necessity of data authentication, and an introduction in Digital Rights Management (DRM) to protect digital multimedia contents is presented. TF techniques provide flexible means to analyze non-stationary audio signals. We have explained the joint TF domain for watermark representation, and have employed pattern recognition schemes for watermark detection. In this chapter; we introduce two watermarking methods; embedding non-linear and linear TF signatures as watermarking signatures. Robustness of the proposed methods against common signal manipulations is also studied in this chapter.

This paper presents a computationally simple and effective methodology based on empirical mode decomposition (EMD) for classification of telephone quality normal and pathological speech signals. EMD is used to decompose continuous normal and pathological speech signals into intrinsic mode functions, which are analyzed to extract physically meaningful and unique temporal and spectral features. Using continuous speech samples from a database of 51 normal and 161 pathological speakers, which has been modified to simulate telephone quality speech under different levels of noise, a linear classifier is used with the feature vector thus obtained to obtain a high classification accuracy, thereby demonstrating the effectiveness of the methodology. The classification accuracy reported in this paper (89.7% for signal-to-noise ratio 30 dB) is a significant improvement over previously reported results for the same task, and demonstrates the utility of our methodology for cost-effective remote voice pathology assessment over telephone channels.