Search

Modern sleep monitoring development is shifting towards the use of unobtrusive sensors combined with algorithms for automatic sleep scoring. Many different combinations of wet and dry electrodes, ear-centered, forehead-mounted or headband-inspired designs have been proposed, alongside an ever growing variety of machine learning algorithms for automatic sleep scoring. Objective: Among candidate positions, those in the facial area and around the ears have the benefit of being relatively hairless, and in our view deserve extra attention. In this paper, we seek to determine the limits to sleep monitoring quality within this spatial constraint. Methods: We compare 13 different, realistic sensor setups derived from the same data set and analysed with the same pipeline. Results: All setups which include both a lateral and an EOG derivation show similar, state-of-the-art performance, with average Cohen’s kappa values of at least 0.80. Conclusion: If large electrode distances are used, positioning is not critical for achieving large sleep-related signal-to-noise-ratio, and hence accurate sleep scoring. Significance: We argue that with the current competitive performance of automated staging approaches, there is a need for establishing an improved benchmark beyond current single human rater scoring. [ABSTRACT FROM AUTHOR]

Using deep neural networks (DNNs) to explore spatial patterns and temporal dynamics of human brain activities has been an important yet challenging problem because the artificial neural networks are hard to be designed manually. There have been several promising deep learning methods, e.g., deep belief network (DBN), convolutional neural network (CNN), and deep sparse recurrent auto-encoder (DSRAE), that can decompose neuroscientific and meaningful spatiotemporal patterns from 4D functional Magnetic Resonance Imaging (fMRI) data. However, those previous studies still depend on hand-crafted neural network architectures and hyperparameters, which are not optimal in various senses. In this paper, we employ the evolutionary algorithms (EA) to optimize the deep neural architecture of DSRAE by minimizing the expected loss of initialized models, named eNAS-DSRAE (evolutionary Neural Architecture Search on Deep Sparse Recurrent Auto-Encoder). Also, validation experiments are designed and performed on the publicly available human connectome project (HCP) 900 datasets, and the results achieved by the optimized eNAS-DSRAE suggested that our framework can successfully identify the spatiotemporal features and perform better than the hand-crafted neural network models. To our best knowledge, the proposed eNAS-DSRAE is not only among the earliest NAS models that can extract connectome-scale meaningful spatiotemporal brain networks from 4D fMRI data, but also is an effective framework to optimize the RNN-based models. [ABSTRACT FROM AUTHOR]

Coronary artery disease (CAD) is a cardiovascular condition with high morbidity and mortality. Intravascular optical coherence tomography (IVOCT) has been considered as an optimal imagining system for the diagnosis and treatment of CAD. Constrained by Nyquist theorem, dense sampling in IVOCT attains high resolving power to delineate cellular structures/features. There is a trade-off between high spatial resolution and fast scanning rate for coronary imaging. In this paper, we propose a viable spectral-spatial acquisition method that down-scales the sampling process in both spectral and spatial domain while maintaining high quality in image reconstruction. The down-scaling schedule boosts data acquisition speed without any hardware modifications. Additionally, we propose a unified multi-scale reconstruction framework, namely Multiscale-Spectral-Spatial-Magnification Network (MSSMN), to resolve highly down-scaled (compressed) OCT images with flexible magnification factors. We incorporate the proposed methods into Spectral Domain OCT (SD-OCT) imaging of human coronary samples with clinical features such as stent and calcified lesions. Our experimental results demonstrate that spectral-spatial down-scaled data can be better reconstructed than data that are down-scaled solely in either spectral or spatial domain. Moreover, we observe better reconstruction performance using MSSMN than using existing reconstruction methods. Our acquisition method and multi-scale reconstruction framework, in combination, may allow faster SD-OCT inspection with high resolution during coronary intervention. [ABSTRACT FROM AUTHOR]

Machine learning techniques used in computer-aided medical image analysis usually suffer from the domain shift problem caused by different distributions between source/reference data and target data. As a promising solution, domain adaptation has attracted considerable attention in recent years. The aim of this paper is to survey the recent advances of domain adaptation methods in medical image analysis. We first present the motivation of introducing domain adaptation techniques to tackle domain heterogeneity issues for medical image analysis. Then we provide a review of recent domain adaptation models in various medical image analysis tasks. We categorize the existing methods into shallow and deep models, and each of them is further divided into supervised, semi-supervised and unsupervised methods. We also provide a brief summary of the benchmark medical image datasets that support current domain adaptation research. This survey will enable researchers to gain a better understanding of the current status, challenges and future directions of this energetic research field. [ABSTRACT FROM AUTHOR]

Breast histology image classification is a crucial step in the early diagnosis of breast cancer. In breast pathological diagnosis, Convolutional Neural Networks (CNNs) have demonstrated great success using digitized histology slides. However, tissue classification is still challenging due to the high visual variability of the large-sized digitized samples and the lack of contextual information. In this paper, we propose a novel CNN, called Multi-level Context and Uncertainty aware (MCUa) dynamic deep learning ensemble model. MCUa model consists of several multi-level context-aware models to learn the spatial dependency between image patches in a layer-wise fashion. It exploits the high sensitivity to the multi-level contextual information using an uncertainty quantification component to accomplish a novel dynamic ensemble model. MCUa model has achieved a high accuracy of 98.11% on a breast cancer histology image dataset. Experimental results show the superior effectiveness of the proposed solution compared to the state-of-the-art histology classification models. [ABSTRACT FROM AUTHOR]

Seizure prediction for drug-refractory epilepsy patients can improve their quality of life, reduce their anxiety, and help them take the necessary precautions. Nowadays, numerous deep learning algorithms have been proposed to predict seizure onset and obtain better performance than traditional machine learning methods. However, these methods require a large set of parameters and large hardware resources; they also have high energy consumption. Therefore, these methods cannot be implemented on compact, low-power wearable, or implantable medical devices. The devices should operate on a real-time basis to continually inform the epileptic patients. In this paper, we describe energy-efficient and hardware-friendly methods to predict the epileptic seizures. A model of only 45 kB was obtained by the neural architecture search and was evaluated across three datasets. The overall sensitivity, false prediction rate, and area under receiver operating characteristic curve were 99.81%, 0.005/h, 1 and 93.48%, 0.063/h, 0.977 and 85.19%, 0.116/h, 0.933, respectively, for three public datasets. This model was further reduced with several methods for model compression. The performances for the model sizes less than 50 kB for scalp EEG data and less than 10 kB for intracranial EEG data outperformed all the other models of similar model sizes. In particular, the energy consumption estimation was less than 10 mJ per inference for scalp EEG signal and less than 0.5 mJ per inference for intracranial EEG signal, which meet the requirements for low-power wearable and implantable devices, respectively. [ABSTRACT FROM AUTHOR]

Objective: To propose a deep-learning network for the diagnosis of two corneal diseases: Fuchs’ endothlelial dystrophy and keratoconus, based on optical coherence tomography (OCT) images of the cornea. Methods: In this paper, we propose a novel network with parallel resolution-specific encoders and composite classification features to directly diagnose Fuchs’ endothelial dystrophy and keratoconus using OCT images. Our proposed network consists of a multi-resolution input, multiple parallel encoders, and a composite of convolutional and dense features for classification. The purpose of using parallel resolution-specific encoders is to perform multi-resolution feature fusion. Also, using composite classification features enhances the dense feature learning. We implemented other related networks for comparison with our network and performed k-fold cross-validation on a dataset of 16,721 OCT images. We used saliency maps and sensitivity analysis to visualize our proposed network. Results: The proposed network outperformed other networks with an image classification accuracy of 0.91 and a scan classification accuracy of 0.94. The visualizations show that our network learned better features than other networks. Significance: The proposed methods can potentially be a step towards the early diagnosis of corneal diseases, which is necessary to prevent their progression, hence, prevent loss of vision. [ABSTRACT FROM AUTHOR]

Objective: Multi-modal functional magnetic resonance imaging (fMRI) can be used to make predictions about individual behavioral and cognitive traits based on brain connectivity networks. Methods: To take advantage of complementary information from multi-modal fMRI, we propose an interpretable multi-modal graph convolutional network (MGCN) model, incorporating both fMRI time series and functional connectivity (FC) between each pair of brain regions. Specifically, our model learns a graph embedding from individual brain networks derived from multi-modal data. A manifold-based regularization term is enforced to consider the relationships of subjects both within and between modalities. Furthermore, we propose the gradient-weighted regression activation mapping (Grad-RAM) and the edge mask learning to interpret the model, which is then used to identify significant cognition-related biomarkers. Results: We validate our MGCN model on the Philadelphia Neurodevelopmental Cohort to predict individual wide range achievement test (WRAT) score. Our model obtains superior predictive performance over GCN with a single modality and other competing approaches. The identified biomarkers are cross-validated from different approaches. Conclusion and Significance: This paper develops a new interpretable graph deep learning framework for cognition prediction, with the potential to overcome the limitations of several current data-fusion models. The results demonstrate the power of MGCN in analyzing multi-modal fMRI and discovering significant biomarkers for human brain studies. [ABSTRACT FROM AUTHOR]

In this work, we propose a novel deep learning reconstruction framework for rapid and accurate reconstruction of 4D flow MRI data. Reconstruction is performed on a slice-by-slice basis by reducing artifacts in zero-filled reconstructed complex images obtained from undersampled k-space. A deep residual attention network FlowRAU-Net is proposed, trained separately for each encoding direction with 2D complex image slices extracted from complex 4D images at each temporal frame and slice position. The network was trained and tested on 4D flow MRI data of aortic valvular flow in 18 human subjects. Performance of the reconstructions was measured in terms of image quality, 3-D velocity vector accuracy, and accuracy in hemodynamic parameters. Reconstruction performance was measured for three different k-space undersamplings and compared with one state of the art compressed sensing reconstruction method and three deep learning-based reconstruction methods. The proposed method outperforms state of the art methods in all performance measures for all three different k-space undersamplings. Hemodynamic parameters such as blood flow rate and peak velocity from the proposed technique show good agreement with reference flow parameters. Visualization of the reconstructed image and velocity magnitude also shows excellent agreement with the fully sampled reference dataset. Moreover, the proposed method is computationally fast. Total 4D flow data (including all slices in space and time) for a subject can be reconstructed in 69 seconds on a single GPU. Although the proposed method has been applied to 4D flow MRI of aortic valvular flows, given a sufficient number of training samples, it should be applicable to other arterial flows. [ABSTRACT FROM AUTHOR]

Objective: Energy Expenditure (EE) estimation plays an important role in objectively evaluating physical activity and its impact on human health. EE during activity can be affected by many factors, including activity intensity, individual physical and physiological characteristics, environment, etc. However, current studies only use very limited information, such as heart rate and step count, to estimate EE, which leads to a low estimation accuracy. Methods: In this study, we proposed a deep multi-branch two-stage regression network (DMTRN) to effectively fuse a variety of related information including motion information, physiological characteristics, and human physical information, which significantly improved the EE estimation accuracy. The proposed DMTRN consists of two main modules: a multi-branch convolutional neural network module which is used to extract multi-scale context features from electrocardiogram (ECG) and inertial measurement unit (IMU) data, and a two-stage regression module which aggregated the extracted multi-scale context features containing the physiological and motion information and the anthropometric features to accurately estimate EE. Results: Experiments performed on 33 participants show that our proposed method is more accurate and the average root mean square error (RMSE) is reduced by 22.8% compared with previous works. Conclusion: The EE estimation accuracy was improved by the proposed DMTRN model with a well-designed network structure and new input signal ECG. Significance: This study verified that ECG was much more effective than HR for EE estimation and cast light on EE estimation using the deep learning method. [ABSTRACT FROM AUTHOR]

Objective: The next generation prosthetic hand that moves and feels like a real hand requires a robust neural interconnection between the human minds and machines. Methods: Here we present a neuroprosthetic system to demonstrate that principle by employing an artificial intelligence (AI) agent to translate the amputee’s movement intent through a peripheral nerve interface. The AI agent is designed based on the recurrent neural network (RNN) and could simultaneously decode six degree-of-freedom (DOF) from multichannel nerve data in real-time. The decoder’s performance is characterized in motor decoding experiments with three human amputees. Results: First, we show the AI agent enables amputees to intuitively control a prosthetic hand with individual finger and wrist movements up to 97-98% accuracy. Second, we demonstrate the AI agent’s real-time performance by measuring the reaction time and information throughput in a hand gesture matching task. Third, we investigate the AI agent’s long-term uses and show the decoder’s robust predictive performance over a 16-month implant duration. Conclusion & significance: Our study demonstrates the potential of AI-enabled nerve technology, underling the next generation of dexterous and intuitive prosthetic hands. [ABSTRACT FROM AUTHOR]

Objective: Statistical shape models have been successfully used in numerous biomedical image analysis applications where prior shape information is helpful such as organ segmentation or data augmentation when training deep learning models. However, training such models requires large data sets, which are often not available and, hence, shape models frequently fail to represent local details of unseen shapes. This work introduces a kernel-based method to alleviate this problem via so-called model localization. It is specifically designed to be used in large-scale shape modeling scenarios like deep learning data augmentation and fits seamlessly into the classical shape modeling framework. Method: Relying on recent advances in multi-level shape model localization via distance-based covariance matrix manipulations and Grassmannian-based level fusion, this work proposes a novel and computationally efficient kernel-based localization technique. Moreover, a novel way to improve the specificity of such models via normalizing flow-based density estimation is presented. Results: The method is evaluated on the publicly available JSRT/SCR chest X-ray and IXI brain data sets. The results confirm the effectiveness of the kernelized formulation and also highlight the models’ improved specificity when utilizing the proposed density estimation method. Conclusion: This work shows that flexible and specific shape models from few training samples can be generated in a computationally efficient way by combining ideas from kernel theory and normalizing flows. Significance: The proposed method together with its publicly available implementation allows to build shape models from few training samples directly usable for applications like data augmentation. [ABSTRACT FROM AUTHOR]

The accurate detection of physiologically-related events in photopletismographic (PPG) and phonocardiographic (PCG) signals, recorded by wearable sensors, is mandatory to perform the estimation of relevant cardiovascular parameters like the heart rate and the blood pressure. However, the measurement performed in uncontrolled conditions without clinical supervision leaves the detection quality particularly susceptible to noise and motion artifacts. This work proposes a new fully-automatic computational framework, based on convolutional networks, to identify and localize fiducial points in time as the foot, maximum slope and peak in PPG signal and the S1 sound in the PCG signal, both acquired by a custom chest sensor, described recently in the literature by our group. The event detection problem was reframed as a single hybrid regression-classification problem entailing a custom neural architecture to process sequentially the PPG and PCG signals. Tests were performed analysing four different acquisition conditions (rest, cycling, rest recovery and walking). Cross-validation results for the three PPG fiducial points showed identification accuracy greater than 93 % and localization error (RMSE) less than 10 ms. As expected, cycling and walking conditions provided worse results than rest and recovery, however reaching an accuracy greater than 90 % and a localization error less than 15 ms. Likewise, the identification and localization error for S1 sound were greater than 90 % and less than 25 ms. Overall, this study showcased the ability of the proposed technique to detect events with high accuracy not only for steady acquisitions but also during subject movements. We also showed that the proposed network outperformed traditional Shannon-energy-envelope method in the detection of S1 sound, reaching detection performance comparable to state of the art algorithms. Therefore, we argue that coupling chest sensors and deep learning processing techniques may disclose wearable devices to unobtrusively acquire health information, being less affected by noise and motion artifacts. [ABSTRACT FROM AUTHOR]

Objective: We show that state-of-the-art deep neural networks achieve superior results in regression-based multi-class proportional myoelectric hand prosthesis control than two common baseline approaches, and we analyze the neural network mapping to explain why this is the case. Methods: Feedforward neural networks and baseline systems are trained on an offline corpus of 11 able-bodied subjects and 4 prosthesis wearers, using the $R^2$ score as metric. Analysis is performed using diverse qualitative and quantitative approaches, followed by a rigorous evaluation. Results: Our best neural networks have at least three hidden layers with at least 128 neurons per layer; smaller architectures, as used by many prior studies, perform substantially worse. The key to good performance is to both optimally regress the target movement, and to suppress spurious movements. Due to the properties of the underlying data, this is impossible to achieve with linear methods, but can be attained with high exactness using sufficiently large neural networks. Conclusion: Neural networks perform significantly better than common linear approaches in the given task, in particular when sufficiently large architectures are used. This can be explained by salient properties of the underlying data, and by theoretical and experimental analysis of the neural network mapping. Significance: To the best of our knowledge, this work is the first one in the field which not only reports that large and deep neural networks are superior to existing architectures, but also explains this result. [ABSTRACT FROM AUTHOR]

Parkinson’s disease (PD) is a chronic, non-reversible neurodegenerative disorder, and freezing of gait (FOG) is one of the most disabling symptoms in PD as it is often the leading cause of falls and injuries that drastically reduces patients’ quality of life. In order to monitor continuously and objectively PD patients who suffer from FOG and enable the possibility of on-demand cueing assistance, a sensor-based FOG detection solution can help clinicians manage the disease and help patients overcome freezing episodes. Many recent studies have leveraged deep learning models to detect FOG using signals extracted from inertial measurement unit (IMU) devices. Usually, the latent features and patterns of FOG are discovered from either the time or frequency domain. In this study, we investigated the use of the time-frequency domain by applying the Continuous Wavelet Transform to signals from IMUs placed on the lower limbs of 63 PD patients who suffered from FOG. We built convolutional neural networks to detect the FOG occurrences, and employed the Bayesian Optimisation approach to obtain the hyper-parameters. The results showed that the proposed subject-independent model was able to achieve a geometric mean of 90.7% and a F1 score of 91.5%. [ABSTRACT FROM AUTHOR]

Objective: Advances in the motor imagery (MI)-based brain-computer interfaces (BCIs) allow control of several applications by decoding neurophysiological phenomena, which are usually recorded by electroencephalography (EEG) using a non-invasive technique. Despite significant advances in MI-based BCI, EEG rhythms are specific to a subject and various changes over time. These issues point to significant challenges to enhance the classification performance, especially in a subject-independent manner. Methods: To overcome these challenges, we propose MIN2Net, a novel end-to-end multi-task learning to tackle this task. We integrate deep metric learning into a multi-task autoencoder to learn a compact and discriminative latent representation from EEG and perform classification simultaneously. Results: This approach reduces the complexity in pre-processing, results in significant performance improvement on EEG classification. Experimental results in a subject-independent manner show that MIN2Net outperforms the state-of-the-art techniques, achieving an F1-score improvement of 6.72% and 2.23% on the SMR-BCI and OpenBMI datasets, respectively. Conclusion: We demonstrate that MIN2Net improves discriminative information in the latent representation. Significance: This study indicates the possibility and practicality of using this model to develop MI-based BCI applications for new users without calibration. [ABSTRACT FROM AUTHOR]

Objective: Automatic detection and analysis of respiratory events in sleep using a single respiratoryeffort belt and deep learning. Methods: Using 9,656 polysomnography recordings from the Massachusetts General Hospital (MGH), we trained a neural network (WaveNet) to detect obstructive apnea, central apnea, hypopnea and respiratory-effort related arousals. Performance evaluation included event-based analysis and apnea-hypopnea index (AHI) stratification. The model was further evaluated on a public dataset, the Sleep-Heart-Health-Study-1, containing 8,455 polysomnographic recordings. Results: For binary apnea event detection in the MGH dataset, the neural network obtained a sensitivity of 68%, a specificity of 98%, a precision of 65%, a F1-score of 67%, and an area under the curve for the receiver operating characteristics curve and precision-recall curve of 0.93 and 0.71, respectively. AHI prediction resulted in a mean difference of 0.41 $\pm$ 7.8 and a r2 of 0.90. For the multiclass task, we obtained varying performances: 84% of all labeled central apneas were correctly classified, whereas this metric was 51% for obstructive apneas, 40% for respiratory effort related arousals and 23% for hypopneas. Conclusion: Our fully automated method can detect respiratory events and assess the AHI accurately. Differentiation of event types is more difficult and may reflect in part the complexity of human respiratory output and some degree of arbitrariness in the criteria used during manual annotation. Significance: The current gold standard of diagnosing sleep-disordered breathing, using polysomnography and manual analysis, is time-consuming, expensive, and only applicable in dedicated clinical environments. Automated analysis using a single effort belt signal overcomes these limitations. [ABSTRACT FROM AUTHOR]

Objective: Brain-computer interfaces (BCI) studies are increasingly leveraging different attributes of multiple signal modalities simultaneously. Bimodal data acquisition protocols combining the temporal resolution of electroencephalography (EEG) with the spatial resolution of functional near-infrared spectroscopy (fNIRS) require novel approaches to decoding. Methods: We present an EEG-fNIRS Hybrid BCI that employs a new bimodal deep neural network architecture consisting of two convolutional sub-networks (subnets) to decode overt and imagined speech. Features from each subnet are fused before further feature extraction and classification. Nineteen participants performed overt and imagined speech in a novel cue-based paradigm enabling investigation of stimulus and linguistic effects on decoding. Results: Using the hybrid approach, classification accuracies (46.31% and 34.29% for overt and imagined speech, respectively (chance: 25%)) indicated a significant improvement on EEG used independently for imagined speech (p = 0.020) while tending towards significance for overt speech (p = 0.098). In comparison with fNIRS, significant improvements for both speech-types were achieved with bimodal decoding (p<0.001). There was a mean difference of ∼12.02% between overt and imagined speech with accuracies as high as 87.18% and 53%. Deeper subnets enhanced performance while stimulus effected overt and imagined speech in significantly different ways. Conclusion: The bimodal approach was a significant improvement on unimodal results for several tasks. Results indicate the potential of multi-modal deep learning for enhancing neural signal decoding. Significance: This novel architecture can be used to enhance speech decoding from bimodal neural signals. [ABSTRACT FROM AUTHOR]

Objective: Graphical deep learning models provide a desirable way for brain functional connectivity analysis. However, the application of current graph deep learning models to brain network analysis is challenging due to the limited sample size and complex relationships between different brain regions. Method: In this work, a graph convolutional network (GCN) based framework is proposed by exploiting the information from both region-to-region connectivities of the brain and subject-subject relationships. We first construct an affinity subject-subject graph followed by GCN analysis. A Laplacian regularization term is introduced in our model to tackle the overfitting problem. We apply and validate the proposed model to the Philadelphia Neurodevelopmental Cohort for the brain cognition study. Results: Experimental analysis shows that our proposed framework outperforms other competing models in classifying groups with low and high Wide Range Achievement Test (WRAT) scores. Moreover, to examine each brain region’s contribution to cognitive function, we use the occlusion sensitivity analysis method to identify cognition-related brain functional networks. The results are consistent with previous research yet yield new findings. Conclusion and significance: Our study demonstrates that GCN incorporating prior knowledge about brain networks offers a powerful way to detect important brain networks and regions associated with cognitive functions. [ABSTRACT FROM AUTHOR]

Objective: Most cardiac arrest patients who are successfully resuscitated are initially comatose due to hypoxic-ischemic brain injury. Quantitative electroencephalography (EEG) provides valuable prognostic information. However, prior approaches largely rely on snapshots of the EEG, without taking advantage of temporal information. Methods: We present a recurrent deep neural network with the goal of capturing temporal dynamics from longitudinal EEG data to predict long-term neurological outcomes. We utilized a large international dataset of continuous EEG recordings from 1,038 cardiac arrest patients from seven hospitals in Europe and the US. Poor outcome was defined as a Cerebral Performance Category (CPC) score of 3-5, and good outcome as CPC score 0-2 at 3 to 6-months after cardiac arrest. Model performance is evaluated using 5-fold cross validation. Results: The proposed approach provides predictions which improve over time, beginning from an area under the receiver operating characteristic curve (AUC-ROC) of 0.78 (95% CI: 0.72-0.81) at 12 hours, and reaching 0.88 (95% CI: 0.85-0.91) by 66 h after cardiac arrest. At 66 h, (sensitivity, specificity) points of interest on the ROC curve for predicting poor outcomes were (32,99)%, (55,95)%, and (62,90)%, (99,23)%, (95,47)%, and (90,62)%; whereas for predicting good outcome, the corresponding operating points were (17,99)%, (47,95)%, (62,90)%, (99,19)%, (95,48)%, (70,90)%. Moreover, the model provides predicted probabilities that closely match the observed frequencies of good and poor outcomes (calibration error 0.04). Conclusions and Significance: These findings suggest that accounting for EEG trend information can substantially improve prediction of neurologic outcomes for patients with coma following cardiac arrest. [ABSTRACT FROM AUTHOR]

Objective: Despitethe proliferation of numerous deep learning methods proposed for generic ECG classification and arrhythmia detection, compact systems with the real-time ability and high accuracy for classifying patient-specific ECG are still few. Particularly, the scarcity of patient-specific data poses an ultimate challenge to any classifier. Recently, compact 1D Convolutional Neural Networks (CNNs) have achieved the state-of-the-art performance level for the accurate classification of ventricular and supraventricular ectopic beats. However, several studies have demonstrated the fact that the learning performance of the conventional CNNs is limited because they are homogenous networks with a basic (linear) neuron model. In order to address this deficiency and further boost the patient-specific ECG classification performance, in this study, we propose 1D Self-organized Operational Neural Networks (1D Self-ONNs). Methods: Due to its self-organization capability, Self-ONNs have the utmost advantage and superiority over conventional ONNs where the prior operator search within the operator set library to find the best possible set of operators is entirely avoided. Results: Under AAMI recommendations and with minimal common training data used, over the entire MIT-BIH dataset 1D Self-ONNs have achieved 98% and 99.04% average accuracies, 76.6% and 93.7% average F1 scores on supra-ventricular and ventricular ectopic beat (VEB) classifications, respectively, which is the highest performance level ever reported. Conclusion: As the first study where 1D Self-ONNs are ever proposed for a classification task, our results over the MIT-BIH arrhythmia benchmark database demonstrate that 1D Self-ONNs can surpass 1D CNNs with a significant margin while having a similar computational complexity. [ABSTRACT FROM AUTHOR]

Heart failure (HF) is one of the most prevalent life-threatening cardiovascular diseases in which 6.5 million people are suffering in the USA and more than 23 million worldwide. Mechanical circulatory support of HF patients can be achieved by implanting a left ventricular assist device (LVAD) into HF patients as a bridge to transplant, recovery or destination therapy and can be controlled by measurement of normal and abnormal pulmonary arterial wedge pressure (PAWP). While there are no commercial long-term implantable pressure sensors to measure PAWP, real-time non-invasive estimation of abnormal and normal PAWP becomes vital. In this work, first an improved Harris Hawks optimizer algorithm called HHO+ is presented and tested on 24 unimodal and multimodal benchmark functions. Second, a novel fully Elman neural network (FENN) is proposed to improve the classification performance. Finally, four novel 18-layer deep learning methods of convolutional neural networks (CNNs) with multi-layer perceptron (CNN-MLP), CNN with Elman neural networks (CNN-ENN), CNN with fully Elman neural networks (CNN-FENN), and CNN with fully Elman neural networks optimized by HHO+ algorithm (CNN-FENN-HHO+) for classification of abnormal and normal PAWP using estimated HVAD pump flow were developed and compared. The estimated pump flow was derived by a non-invasive method embedded into the commercial HVAD controller. The proposed methods are evaluated on an imbalanced clinical dataset using 5-fold cross-validation. The proposed CNN-FENN-HHO+ method outperforms the proposed CNN-MLP, CNN-ENN and CNN-FENN methods and improved the classification performance metrics across 5-fold cross-validation with an average sensitivity of 79%, accuracy of 78% and specificity of 76%. The proposed methods can reduce the likelihood of hazardous events like pulmonary congestion and ventricular suction for HF patients and notify identified abnormal cases to the hospital, clinician and cardiologist for emergency action, which can diminish the mortality rate of patients with HF. [ABSTRACT FROM AUTHOR]

Objective: Target identification in brain-computer interface (BCI) spellers refers to the electroencephalogram (EEG) classification for predicting the target character that the subject intends to spell. When the visual stimulus of each character is tagged with a distinct frequency, the EEG records steady-state visually evoked potentials (SSVEP) whose spectrum is dominated by the harmonics of the target frequency. In this setting, we address the target identification and propose a novel deep neural network (DNN) architecture. Method: The proposed DNN processes the multi-channel SSVEP with convolutions across the sub-bands of harmonics, channels, time, and classifies at the fully connected layer. We test with two publicly available large scale (the benchmark and BETA) datasets consisting of in total 105 subjects with 40 characters. Our first stage training learns a global model by exploiting the statistical commonalities among all subjects, and the second stage fine tunes to each subject separately by exploiting the individualities. Results: Our DNN achieves impressive information transfer rates (ITRs) on both datasets, 265.23 bits/min and 196.59 bits/min, respectively, with only 0.4 seconds of stimulation. The code is available for reproducibility at https://github.com/osmanberke/Deep-SSVEP-BCI. Conclusion: The presented DNN strongly outperforms the state-of-the-art techniques as our accuracy and ITR rates are the highest ever reported performance results on these datasets. Significance: Due to its unprecedentedly high speller ITRs and flawless applicability to general SSVEP systems, our technique has great potential in various biomedical engineering settings of BCIs such as communication, rehabilitation and control. [ABSTRACT FROM AUTHOR]

Objective: Noise and disturbances hinder effective interpretation of recorded ECG. To identify the clean parts of a recording, free from such disturbances, various quality indicators have been developed. Previous instances of these indicators focus on human-defined desirable properties of a clean signal. The reliance on human-specified properties places an inherent limitation on the potential power of signal quality indicators. To move away from this limitation, we propose a data-driven quality indicator. Methods: We use an unsupervised deep learning model, the auto-encoder, to derive the quality indicator. For different quality assessment settings we compare the performance of our quality indicator with traditional indicators. Results: The data-driven method performs consistently strong across tasks while performance of the traditional indicators varies strongly from task to task. Conclusion: This strong performance indicates the potential of data-driven quality indicators for use in ECG processing, removing the reliance on expert-specified desirable properties. Significance: The proposed methodology can easily be extended towards learning quality indicators in other data modalities. [ABSTRACT FROM AUTHOR]

Objective: Many recent studies suggest that brain deformation resulting from head impacts are linked to the corresponding clinical outcome, such as mild traumatic brain injury (mTBI). Even if several finite element (FE) head models have been developed and validated to calculate brain deformation based on impact kinematics, the clinical application of these FE head models is limited due to the time-consuming nature of FE simulations. This work aims to accelerate the brain deformation calculation and thus improve the potential for clinical applications. Methods: We propose a deep learning head model with a five-layer deep neural network and feature engineering, and trained and tested the model on 2511 head impacts from a combination of head model simulations and on-field college football and mixed martial arts impacts. Results: The proposed deep learning head model can calculate the maximum principal strain (Green Lagrange) for every element in the entire brain in less than 0.001 s with an average root mean squared error of 0.022 and a standard deviation of 0.001 over twenty repeats with random data partition and model initialization. Conclusion: Trained and tested using the dataset of 2511 head impacts, this model can be applied to various sports in the calculation of brain strain with accuracy, and its applicability can even further be extended by incorporating data from other types of head impacts. Significance: In addition to the potential clinical application in real-time brain deformation monitoring, this model will help researchers estimate the brain strain from a large number of head impacts more efficiently than using FE models. [ABSTRACT FROM AUTHOR]

Objective: A large number of atrial fibrillation (AF) detectors have been published in recent years, signifying that the comparison of detector performance plays a central role, though not always consistent. The aim of this study is to shed needed light on aspects crucial to the evaluation of detection performance. Methods: Three types of AF detector, using either information on rhythm, rhythm and morphology, or segments of ECG samples, are implemented and studied on both real and simulated ECG signals. The properties of different performance measures are investigated, for example, in relation to dataset imbalance. Results: The results show that performance can differ considerably depending on the way detector output is compared to database annotations, i.e., beat-to-beat, segment-to-segment, or episode-to-episode comparison. Moreover, depending on the type of detector, the results substantiate that physiological and technical factors, e.g., changes in ECG morphology, rate of atrial premature beats, and noise level, can have a considerable influence on performance. Conclusion: The present study demonstrates overall strengths and weaknesses of different types of detector, highlights challenges in AF detection, and proposes five recommendations on how to handle data and characterize performance. [ABSTRACT FROM AUTHOR]

Left ventricular assist devices (LVADs) are mechanical pumps, which can be used to support heart failure (HF) patients as bridge to transplant and destination therapy. To automatically adjust the LVAD speed, a physiological control system needs to be designed to respond to variations of patient hemodynamics across a variety of clinical scenarios. These control systems require pressure feedback signals from the cardiovascular system. However, there are no suitable long-term implantable sensors available. In this study, a novel real-time deep convolutional neural network (CNN) for estimation of preload based on the LVAD flow was proposed. A new sensorless adaptive physiological control system for an LVAD pump was developed using the full dynamic form of model free adaptive control (FFDL-MFAC) and the proposed preload estimator to maintain the patient conditions in safe physiological ranges. The CNN model for preload estimation was trained and evaluated through 10-fold cross validation on 100 different patient conditions and the proposed sensorless control system was assessed on a new testing set of 30 different patient conditions across six different patient scenarios. The proposed preload estimator was extremely accurate with a correlation coefficient of 0.97, root mean squared error of 0.84 mmHg, reproducibility coefficient of 1.56 mmHg, coefficient of variation of 14.44%, and bias of 0.29 mmHg for the testing dataset. The results also indicate that the proposed sensorless physiological controller works similarly to the preload-based physiological control system for LVAD using measured preload to prevent ventricular suction and pulmonary congestion. This study shows that the LVADs can respond appropriately to changing patient states and physiological demands without the need for additional pressure or flow measurements. [ABSTRACT FROM AUTHOR]

Objective: This article shows the interest in deep learning techniques to detect motor imagery (MI) from raw electroencephalographic (EEG) signals when a functional electrical stimulation is added or not. Impacts of electrode montages and bandwidth are also reported. The perspective of this work is to improve the detection of intraoperative awareness during general anesthesia. Methods: Various architectures of EEGNet were investigated to optimize MI detection. They have been compared to the state-of-the-art classifiers in Brain-Computer Interfaces (based on Riemannian geometry, linear discriminant analysis), and other deep learning architectures (deep convolution network, shallow convolutional network). EEG data were measured from 22 participants performing motor imagery with and without median nerve stimulation. Results: The proposed architecture of EEGNet reaches the best classification accuracy (83.2%) and false-positive rate (FPR 19.0%) for a setup with only six electrodes over the motor cortex and frontal lobe and for an extended 4-38 Hz EEG frequency range while the subject is being stimulated via a median nerve. Configurations with a larger number of electrodes result in higher accuracy (94.5%) and FPR (6.1%) for 128 electrodes (and respectively 88.0% and 12.9% for 13 electrodes). Conclusion: The present work demonstrates that using an extended EEG frequency band and a modified EEGNet deep neural network increases the accuracy of MI detection when used with as few as 6 electrodes which include frontal channels. Significance: The proposed method contributes to the development of Brain-Computer Interface systems based on MI detection from EEG. [ABSTRACT FROM AUTHOR]

Objectives: Evaluation of hypernasality requires extensive perceptual training by clinicians and extending this training on a large scale internationally is untenable; this compounds the health disparities that already exist among children with cleft. In this work, we present the objective hypernasality measure (OHM), a speech-based algorithm that automatically measures hypernasality in speech, and validate it relative to a group of trained clinicians. Methods: We trained a deep neural network (DNN) on approximately 100 hours of a publicly-available healthy speech corpus to detect the presence of nasal acoustic cues generated through the production of nasal consonants and nasalized phonemes in speech. Importantly, this model does not require any clinical data for training. The posterior probabilities of the deep learning model were aggregated at the sentence and speaker-levels to compute the OHM. Results: The results showed that the OHM was significantly correlated with perceptual hypernasality ratings from the Americleft database (r = 0.797, p < 0.001) and the New Mexico Cleft Palate Center (NMCPC) database (r = 0.713, p < 0.001). In addition, we evaluated the relationship between the OHM and articulation errors; the sensitivity of the OHM in detecting the presence of very mild hypernasality; and established the internal reliability of the metric. Further, the performance of the OHM was compared with a DNN regression algorithm directly trained on the hypernasal speech samples. Significance: The results indicate that the OHM is able to measure the severity of hypernasality on par with Americleft-trained clinicians on thisdataset. [ABSTRACT FROM AUTHOR]