Parkinson's Disease EMG Data Augmentation and Simulation with DCGANs and Style Transfer

EEGNet classification of sleep EEG for individual specialization based on data augmentation

Sleep is an essential part of human life, and the quality of one's sleep is also an important indicator of one's health. Analyzing the Electroencephalogram (EEG) signals of a person during sleep makes it possible to understand the sleep status and give relevant rest or medical advice. In this paper, a decent amount of artificial data generated with a data augmentation method based on Discrete Cosine Transform from a small amount of real experimental data of a specific individual is introduced. A classification model with an accuracy of 92.85% has been obtained. By mixing the data augmentation with the public database and training with the EEGNet, we obtained a classification model with significantly higher accuracy for the specific individual. The experiments have demonstrated that we can circumvent the subject-independent problem in sleep EEG in this way and use only a small amount of labeled data to customize a dedicated classification model with high accuracy.

Deep learning and predictive modelling for generating normalised muscle function parameters from signal images of mandibular electromyography

Challenges arise in accessing archived signal outputs due to proprietary software limitations. There is a notable lack of exploration in open-source mandibular EMG signal conversion for continuous access and analysis, hindering tasks such as pattern recognition and predictive modelling for temporomandibular joint complex function. To Develop a workflow to extract normalised signal parameters from images of mandibular muscle EMG and identify optimal clustering methods for quantifying signal intensity and activity durations. A workflow utilising OpenCV, variational encoders and Neurokit2 generated and augmented 866 unique EMG signals from jaw movement exercises. k-means, GMM and DBSCAN were employed for normalisation and cluster-centric signal processing. The workflow was validated with data collected from 66 participants, measuring temporalis, masseter and digastric muscles. DBSCAN (0.35 to 0.54) and GMM (0.09 to 0.24) exhibited lower silhouette scores for mouth opening, anterior protrusion and lateral excursions, while K-means performed best (0.10 to 0.11) for temporalis and masseter muscles during chewing activities. The current study successfully developed a deep learning workflow capable of extracting normalised signal data from EMG images and generating quantifiable parameters for muscle activity duration and general functional intensity.

A systematic literature review: deep learning techniques for synthetic medical image generation and their applications in radiotherapy

The aim of this systematic review is to determine whether Deep Learning (DL) algorithms can provide a clinically feasible alternative to classic algorithms for synthetic Computer Tomography (sCT). The following categories are presented in this study: ∙ MR-based treatment planning and synthetic CT generation techniques. ∙ Generation of synthetic CT images based on Cone Beam CT images. ∙ Low-dose CT to High-dose CT generation. ∙ Attenuation correction for PET images. To perform appropriate database searches, we reviewed journal articles published between January 2018 and June 2023. Current methodology, study strategies, and results with relevant clinical applications were analyzed as we outlined the state-of-the-art of deep learning based approaches to inter-modality and intra-modality image synthesis. This was accomplished by contrasting the provided methodologies with traditional research approaches. The key contributions of each category were highlighted, specific challenges were identified, and accomplishments were summarized. As a final step, the statistics of all the cited works from various aspects were analyzed, which revealed that DL-based sCTs have achieved considerable popularity, while also showing the potential of this technology. In order to assess the clinical readiness of the presented methods, we examined the current status of DL-based sCT generation.

Synthesis of sEMG Signals for Hand Gestures Using a 1DDCGAN

The emergence of modern prosthetics controlled by bio-signals has been facilitated by AI and microchip technology innovations. AI algorithms are trained using sEMG produced by muscles during contractions. The data acquisition procedure may result in discomfort and fatigue, particularly for amputees. Furthermore, prosthetic companies restrict sEMG signal exchange, limiting data-driven research and reproducibility. GANs present a viable solution to the aforementioned concerns. GANs can generate high-quality sEMG, which can be utilised for data augmentation, decrease the training time required by prosthetic users, enhance classification accuracy and ensure research reproducibility. This research proposes the utilisation of a one-dimensional deep convolutional GAN (1DDCGAN) to generate the sEMG of hand gestures. This approach involves the incorporation of dynamic time wrapping, fast Fourier transform and wavelets as discriminator inputs. Two datasets were utilised to validate the methodology, where five windows and increments were utilised to extract features to evaluate the synthesised sEMG quality. In addition to the traditional classification and augmentation metrics, two novel metrics-the Mantel test and the classifier two-sample test-were used for evaluation. The 1DDCGAN preserved the inter-feature correlations and generated high-quality signals, which resembled the original data. Additionally, the classification accuracy improved by an average of 1.21-5%.

Graph Neural Networks for Parkinson's Disease Monitoring and Alerting

Graph neural networks (GNNs) have been increasingly employed in the field of Parkinson's disease (PD) research. The use of GNNs provides a promising approach to address the complex relationship between various clinical and non-clinical factors that contribute to the progression of PD. This review paper aims to provide a comprehensive overview of the state-of-the-art research that is using GNNs for PD. It presents PD and the motivation behind using GNNs in this field. Background knowledge on the topic is also presented. Our research methodology is based on PRISMA, presenting a comprehensive overview of the current solutions using GNNs for PD, including the various types of GNNs employed and the results obtained. In addition, we discuss open issues and challenges that highlight the limitations of current GNN-based approaches and identify potential paths for future research. Finally, a new approach proposed in this paper presents the integration of new tasks for the engineering of GNNs for PD monitoring and alert solutions.

FoGGAN: Generating Realistic Parkinson's Disease Freezing of Gait Data Using GANs

Data scarcity in the healthcare domain is a major drawback for most state-of-the-art technologies engaging artificial intelligence. The unavailability of quality data due to both the difficulty to gather and label them as well as due to their sensitive nature create a breeding ground for data augmentation solutions. Parkinson's Disease (PD) which can have a wide range of symptoms including motor impairments consists of a very challenging case for quality data acquisition. Generative Adversarial Networks (GANs) can help alleviate such data availability issues. In this light, this study focuses on a data augmentation solution engaging Generative Adversarial Networks (GANs) using a freezing of gait (FoG) symptom dataset as input. The data generated by the so-called FoGGAN architecture presented in this study are almost identical to the original as concluded by a variety of similarity metrics. This highlights the significance of such solutions as they can provide credible synthetically generated data which can be utilized as training dataset inputs to AI applications. Additionally, a DNN classifier's performance is evaluated using three different evaluation datasets and the accuracy results were quite encouraging, highlighting that the FOGGAN solution could lead to the alleviation of the data shortage matter.

Decoding Mental Effort in a Quasi-Realistic Scenario: A Feasibility Study on Multimodal Data Fusion and Classification

Humans' performance varies due to the mental resources that are available to successfully pursue a task. To monitor users' current cognitive resources in naturalistic scenarios, it is essential to not only measure demands induced by the task itself but also consider situational and environmental influences. We conducted a multimodal study with 18 participants (nine female, M = 25.9 with SD = 3.8 years). In this study, we recorded respiratory, ocular, cardiac, and brain activity using functional near-infrared spectroscopy (fNIRS) while participants performed an adapted version of the warship commander task with concurrent emotional speech distraction. We tested the feasibility of decoding the experienced mental effort with a multimodal machine learning architecture. The architecture comprised feature engineering, model optimisation, and model selection to combine multimodal measurements in a cross-subject classification. Our approach reduces possible overfitting and reliably distinguishes two different levels of mental effort. These findings contribute to the prediction of different states of mental effort and pave the way toward generalised state monitoring across individuals in realistic applications.

De novo distillation of thermodynamic affinity from deep learning regulatory sequence models of in vivo protein-DNA binding

Transcription factors (TF) are proteins that bind DNA in a sequence-specific manner to regulate gene transcription. Despite their unique intrinsic sequence preferences, in vivo genomic occupancy profiles of TFs differ across cellular contexts. Hence, deciphering the sequence determinants of TF binding, both intrinsic and context-specific, is essential to understand gene regulation and the impact of regulatory, non-coding genetic variation. Biophysical models trained on in vitro TF binding assays can estimate intrinsic affinity landscapes and predict occupancy based on TF concentration and affinity. However, these models cannot adequately explain context-specific, in vivo binding profiles. Conversely, deep learning models, trained on in vivo TF binding assays, effectively predict and explain genomic occupancy profiles as a function of complex regulatory sequence syntax, albeit without a clear biophysical interpretation. To reconcile these complementary models of in vitro and in vivo TF binding, we developed Affinity Distillation (AD), a method that extracts thermodynamic affinities de-novo from deep learning models of TF chromatin immunoprecipitation (ChIP) experiments by marginalizing away the influence of genomic sequence context. Applied to neural networks modeling diverse classes of yeast and mammalian TFs, AD predicts energetic impacts of sequence variation within and surrounding motifs on TF binding as measured by diverse in vitro assays with superior dynamic range and accuracy compared to motif-based methods. Furthermore, AD can accurately discern affinities of TF paralogs. Our results highlight thermodynamic affinity as a key determinant of in vivo binding, suggest that deep learning models of in vivo binding implicitly learn high-resolution affinity landscapes, and show that these affinities can be successfully distilled using AD. This new biophysical interpretation of deep learning models enables high-throughput in silico experiments to explore the influence of sequence context and variation on both intrinsic affinity and in vivo occupancy.

A myoelectric digital twin for fast and realistic modelling in deep learning

Muscle electrophysiology has emerged as a powerful tool to drive human machine interfaces, with many new recent applications outside the traditional clinical domains, such as robotics and virtual reality. However, more sophisticated, functional, and robust decoding algorithms are required to meet the fine control requirements of these applications. Deep learning has shown high potential in meeting these demands, but requires a large amount of high-quality annotated data, which is expensive and time-consuming to acquire. Data augmentation using simulations, a strategy applied in other deep learning applications, has never been attempted in electromyography due to the absence of computationally efficient models. We introduce a concept of Myoelectric Digital Twin - highly realistic and fast computational model tailored for the training of deep learning algorithms. It enables simulation of arbitrary large and perfectly annotated datasets of realistic electromyography signals, allowing new approaches to muscular signal decoding, accelerating the development of human-machine interfaces.

Investigation of Phase Shifts Using AUC Diagrams: Application to Differential Diagnosis of Parkinson's Disease and Essential Tremor

This study was motivated by the well-known problem of the differential diagnosis of Parkinson's disease and essential tremor using the phase shift between the tremor signals in the antagonist muscles of patients. Different phase shifts are typical for different diseases; however, it remains unclear how this parameter can be used for clinical diagnosis. Neurophysiological papers have reported different estimations of the accuracy of this parameter, which varies from insufficient to 100%. To address this issue, we developed special types of area under the ROC curve (AUC) diagrams and used them to analyze the phase shift. Different phase estimations, including the Hilbert instantaneous phase and the cross-wavelet spectrum mean phase, were applied. The results of the investigation of the clinical data revealed several regularities with opposite directions in the phase shift of the electromyographic signals in patients with Parkinson's disease and essential tremor. The detected regularities provide insights into the contradictory results reported in the literature. Moreover, the developed AUC diagrams show the potential for the investigation of neurodegenerative diseases related to the hyperkinetic movements of the extremities and the creation of high-accuracy methods of clinical diagnosis.

Deep Convolutional Generative Adversarial Network-Based EMG Data Enhancement for Hand Motion Classification

The acquisition of bio-signal from the human body requires a strict experimental setup and ethical approvements, which leads to limited data for the training of classifiers in the era of big data. It will change the situation if synthetic data can be generated based on real data. This article proposes such a kind of multiple channel electromyography (EMG) data enhancement method using a deep convolutional generative adversarial network (DCGAN). The generation procedure is as follows: First, the multiple channels of EMG signals within sliding windows are converted to grayscale images through matrix transformation, normalization, and histogram equalization. Second, the grayscale images of each class are used to train DCGAN so that synthetic grayscale images of each class can be generated with the input of random noises. To evaluate whether the synthetic data own the similarity and diversity with the real data, the classification accuracy index is adopted in this article. A public EMG dataset (that is, ISR Myo-I) for hand motion recognition is used to prove the usability of the proposed method. The experimental results show that adding synthetic data to the training data has little effect on the classification performance, indicating the similarity between real data and synthetic data. Moreover, it is also noted that the average accuracy (five classes) is slightly increased by 1%–2% for support vector machine (SVM) and random forest (RF), respectively, with additional synthetic data for training. Although the improvement is not statistically significant, it implies that the generated data by DCGAN own its new characteristics, and it is possible to enrich the diversity of the training dataset. In addition, cross-validation analysis shows that the synthetic samples have large inter-class distance, reflected by higher cross-validation accuracy of pure synthetic sample classification. Furthermore, this article also demonstrates that histogram equalization can significantly improve the performance of EMG-based hand motion recognition.

A novel sEMG data augmentation based on WGAN-GP

The classification of sEMG signals is fundamental in applications that use mechanical prostheses, making it necessary to work with generalist databases that improve the accuracy of those classifications. Therefore, synthetic signal generation can be beneficial in enriching a database to make it more generalist. This work proposes using a variant of generative adversarial networks to produce synthetic biosignals of sEMG. A convolutional neural network (CNN) was used to classify the movements. The results showed good performance with an increase of 4.07% in a set of movement classification accuracy when 200 synthetic samples were included for each movement. We compared our results to other methodologies, such as Magnitude Warping and Scaling. Both methodologies did not have the same performance in the classification.

Using a multi-head, convolutional neural network with data augmentation to improve electropherogram classification performance

DNA profiles are generated in forensic biology laboratories around the world. It is possible that these profiles are assessed by two independent people in order for the profiles to be 'read'. Recent work has been carried out to develop a neural network model to classify fluorescence in a DNA profile electropherogram and potentially replace one, or both human readers. The ability to use neural networks for this function has been programmed into the software FaSTR™ DNA, which has been validated for use in at least one laboratory in Australia. The work that previously developed a neural network system had a number of limitations, specifically it was computer intensive, did not make the best use of available data, and consequently the performance of this model was sub-optimal in some conditions (particularly for low-intensity peaks). In the current work a new neural network model is developed that makes various improvements on the old model, by using convolutional layers, a multi-head architecture and data augmentation. Results indicate that an improved performance can be expected for low-intensity profiles.

Detecting Task Difficulty of Learners in Colonoscopy: Evidence from Eye-Tracking

Eye-tracking can help decode the intricate control mechanism in human performance. In healthcare, physicians-in-training require extensive practice to improve their healthcare skills. When a trainee encounters any difficulty in the practice, they will need feedback from experts to improve their performance. Personal feedback is time-consuming and subjected to bias. In this study, we tracked the eye movements of trainees during their colonoscopic performance in simulation. We examined changes in eye movement behavior during the moments of navigation loss (MNL), a signature sign for task difficulty during colonoscopy, and tested whether deep learning algorithms can detect the MNL by feeding data from eye-tracking. Human eye gaze and pupil characteristics were learned and verified by the deep convolutional generative adversarial networks (DCGANs); the generated data were fed to the Long Short-Term Memory (LSTM) networks with three different data feeding strategies to classify MNLs from the entire colonoscopic procedure. Outputs from deep learning were compared to the expert's judgment on the MNLs based on colonoscopic videos. The best classification outcome was achieved when we fed human eye data with 1000 synthesized eye data, where accuracy (91.80%), sensitivity (90.91%), and specificity (94.12%) were optimized. This study built an important foundation for our work of developing an education system for training healthcare skills using simulation.

A deep learning based framework for the registration of three dimensional multi-modal medical images of the head

Image registration is a fundamental task in image analysis in which the transform that moves the coordinate system of one image to another is calculated. Registration of multi-modal medical images has important implications for clinical diagnosis, treatment planning, and image-guided surgery as it provides the means of bringing together complimentary information obtained from different image modalities. However, since different image modalities have different properties due to their different acquisition methods, it remains a challenging task to find a fast and accurate match between multi-modal images. Furthermore, due to reasons such as ethical issues and need for human expert intervention, it is difficult to collect a large database of labelled multi-modal medical images. In addition, manual input is required to determine the fixed and moving images as input to registration algorithms. In this paper, we address these issues and introduce a registration framework that (1) creates synthetic data to augment existing datasets, (2) generates ground truth data to be used in the training and testing of algorithms, (3) registers (using a combination of deep learning and conventional machine learning methods) multi-modal images in an accurate and fast manner, and (4) automatically classifies the image modality so that the process of registration can be fully automated. We validate the performance of the proposed framework on CT and MRI images of the head obtained from a publicly available registration database.

Efficient Learning of Healthcare Data from IoT Devices by Edge Convolution Neural Networks

Wireless and mobile health applications promote the development of smart healthcare. Effective diagnosis and feedbacks of remote health data pose significant challenges due to streaming data, high noise, network latency and user privacy. Therefore, we explore efficient edge and cloud design to maintain electrocardiogram classification performance while reducing the communication cost. These contributions include: (1) We introduce a hybrid smart medical architecture named edge convolutional neural networks (EdgeCNN) that balances the capability of edge and cloud computing to address the issue for agile learning of healthcare data from IoT devices. (2) We present an effective deep learning model for electrocardiogram (ECG) inference, which can be deployed to run on edge smart devices for low-latency diagnosis. (3) We design a data enhancement method for ECG based on deep convolutional generative adversarial network to expand ECG data volume. (4) We carried out experiments on two representative datasets to evaluate the effectiveness of the deep learning model of ECG classification based on EdgeCNN. EdgeCNN shows superior to traditional cloud medical systems in terms of network Input/Output (I/O) pressure, architecture cost and system high availability. The deep learning model not only ensures high diagnostic accuracy, but also has advantages in aspect of inference time, storage, running memory and power consumption.

Assisting Barrett's esophagus identification using endoscopic data augmentation based on Generative Adversarial Networks

Barrett's esophagus figured a swift rise in the number of cases in the past years. Although traditional diagnosis methods offered a vital role in early-stage treatment, they are generally time- and resource-consuming. In this context, computer-aided approaches for automatic diagnosis emerged in the literature since early detection is intrinsically related to remission probabilities. However, they still suffer from drawbacks because of the lack of available data for machine learning purposes, thus implying reduced recognition rates. This work introduces Generative Adversarial Networks to generate high-quality endoscopic images, thereby identifying Barrett's esophagus and adenocarcinoma more precisely. Further, Convolution Neural Networks are used for feature extraction and classification purposes. The proposed approach is validated over two datasets of endoscopic images, with the experiments conducted over the full and patch-split images. The application of Deep Convolutional Generative Adversarial Networks for the data augmentation step and LeNet-5 and AlexNet for the classification step allowed us to validate the proposed methodology over an extensive set of datasets (based on original and augmented sets), reaching results of 90% of accuracy for the patch-based approach and 85% for the image-based approach. Both results are based on augmented datasets and are statistically different from the ones obtained in the original datasets of the same kind. Moreover, the impact of data augmentation was evaluated in the context of image description and classification, and the results obtained using synthetic images outperformed the ones over the original datasets, as well as other recent approaches from the literature. Such results suggest promising insights related to the importance of proper data for the accurate classification concerning computer-assisted Barrett's esophagus and adenocarcinoma detection.

Data Augmentation of Surface Electromyography for Hand Gesture Recognition

The range of applications of electromyography-based gesture recognition has increased over the last years. A common problem regularly encountered in literature is the inadequate data availability. Data augmentation, which aims at generating new synthetic data from the existing ones, is the most common approach to deal with this data shortage in other research domains. In the case of surface electromyography (sEMG) signals, there is limited research in augmentation methods and quite regularly the results differ between available studies. In this work, we provide a detailed evaluation of existing (i.e., additive noise, overlapping windows) and novel (i.e., magnitude warping, wavelet decomposition, synthetic sEMG models) strategies of data augmentation for electromyography signals. A set of metrics (i.e., classification accuracy, silhouette score, and Davies-Bouldin index) and visualizations help with the assessment and provides insights about their performance. Methods like signal magnitude warping and wavelet decomposition yield considerable increase (up to 16%) in classification accuracy across two benchmark datasets. Particularly, a significant improvement of 1% in the classification accuracy of the state-of-the-art model in hand gesture recognition is achieved.

Data Augmentation for Motor Imagery Signal Classification Based on a Hybrid Neural Network

As an important paradigm of spontaneous brain-computer interfaces (BCIs), motor imagery (MI) has been widely used in the fields of neurological rehabilitation and robot control. Recently, researchers have proposed various methods for feature extraction and classification based on MI signals. The decoding model based on deep neural networks (DNNs) has attracted significant attention in the field of MI signal processing. Due to the strict requirements for subjects and experimental environments, it is difficult to collect large-scale and high-quality electroencephalogram (EEG) data. However, the performance of a deep learning model depends directly on the size of the datasets. Therefore, the decoding of MI-EEG signals based on a DNN has proven highly challenging in practice. Based on this, we investigated the performance of different data augmentation (DA) methods for the classification of MI data using a DNN. First, we transformed the time series signals into spectrogram images using a short-time Fourier transform (STFT). Then, we evaluated and compared the performance of different DA methods for this spectrogram data. Next, we developed a convolutional neural network (CNN) to classify the MI signals and compared the classification performance of after DA. The Fréchet inception distance (FID) was used to evaluate the quality of the generated data (GD) and the classification accuracy, and mean kappa values were used to explore the best CNN-DA method. In addition, analysis of variance (ANOVA) and paired t-tests were used to assess the significance of the results. The results showed that the deep convolutional generative adversarial network (DCGAN) provided better augmentation performance than traditional DA methods: geometric transformation (GT), autoencoder (AE), and variational autoencoder (VAE) (p < 0.01). Public datasets of the BCI competition IV (datasets 1 and 2b) were used to verify the classification performance. Improvements in the classification accuracies of 17% and 21% (p < 0.01) were observed after DA for the two datasets. In addition, the hybrid network CNN-DCGAN outperformed the other classification methods, with average kappa values of 0.564 and 0.677 for the two datasets.