Evaluating Convolutional Neural Networks as a Method of EEG–EMG Fusion

Wearable robotic exoskeletons have emerged as an exciting new treatment tool for disorders affecting mobility; however, the human–machine interface, used by the patient for device control, requires further improvement before robotic assistance and rehabilitation can be widely adopted. One method, made possible through advancements in machine learning technology, is the use of bioelectrical signals, such as electroencephalography (EEG) and electromyography (EMG), to classify the user's actions and intentions. While classification using these signals has been demonstrated for many relevant control tasks, such as motion intention detection and gesture recognition, challenges in decoding the bioelectrical signals have caused researchers to seek methods for improving the accuracy of these models. One such method is the use of EEG–EMG fusion, creating a classification model that decodes information from both EEG and EMG signals simultaneously to increase the amount of available information. So far, EEG–EMG fusion has been implemented using traditional machine learning methods that rely on manual feature extraction; however, new machine learning methods have emerged that can automatically extract relevant information from a dataset, which may prove beneficial during EEG–EMG fusion. In this study, Convolutional Neural Network (CNN) models were developed using combined EEG–EMG inputs to determine if they have potential as a method of EEG–EMG fusion that automatically extracts relevant information from both signals simultaneously. EEG and EMG signals were recorded during elbow flexion–extension and used to develop CNN models based on time–frequency (spectrogram) and time (filtered signal) domain image inputs. The results show a mean accuracy of 80.51 ± 8.07% for a three-class output (33.33% chance level), with an F-score of 80.74%, using time–frequency domain-based models. This work demonstrates the viability of CNNs as a new method of EEG–EMG fusion and evaluates different signal representations to determine the best implementation of a combined EEG–EMG CNN. It leverages modern machine learning methods to advance EEG–EMG fusion, which will ultimately lead to improvements in the usability of wearable robotic exoskeletons.

Application of Dense Neural Networks for Detection of Atrial Fibrillation and Ranking of Augmented ECG Feature Set

Considering the significant burden to patients and healthcare systems globally related to atrial fibrillation (AF) complications, the early AF diagnosis is of crucial importance. In the view of prominent perspectives for fast and accurate point-of-care arrhythmia detection, our study optimizes an artificial neural network (NN) classifier and ranks the importance of enhanced 137 diagnostic ECG features computed from time and frequency ECG signal representations of short single-lead strips available in 2017 Physionet/CinC Challenge database. Based on hyperparameters’ grid search of densely connected NN layers, we derive the optimal topology with three layers and 128, 32, 4 neurons per layer (DenseNet-3@128-32-4), which presents maximal F1-scores for classification of Normal rhythms (0.883, 5076 strips), AF (0.825, 758 strips), Other rhythms (0.705, 2415 strips), Noise (0.618, 279 strips) and total F1 relevant to the CinC Challenge of 0.804, derived by five-fold cross-validation. DenseNet-3@128-32-4 performs equally well with 137 to 32 features and presents tolerable reduction by about 0.03 to 0.06 points for limited input sets, including 8 and 16 features, respectively. The feature reduction is linked to effective application of a comprehensive method for computation of the feature map importance based on the weights of the activated neurons through the total path from input to specific output in DenseNet. The detailed analysis of 20 top-ranked ECG features with greatest importance to the detection of each rhythm and overall of all rhythms reveals DenseNet decision-making process, noticeably corresponding to the cardiologists’ diagnostic point of view.

Using the Kullback-Leibler Divergence and Kolmogorov-Smirnov Test to Select Input Sizes to the Fault Diagnosis Problem Based on a CNN Model

Choosing a suitable size for signal representations, e.g., frequency spectra, in a given machine learning problem is not a trivial task. It may strongly affect the performance of the trained models. Many solutions have been proposed to solve this problem. Most of them rely on designing an optimized input or selecting the most suitable input according to an exhaustive search. In this work, we used the Kullback-Leibler Divergence and the Kolmogorov-Smirnov Test to measure the dissimilarity among signal representations belonging to equal and different classes, i.e., we measured the intraclass and interclass dissimilarities. Moreover, we analyzed how this information relates to the classifier performance. The results suggested that both the interclass and intraclass dissimilarities were related to the model accuracy since they indicate how easy a model can learn discriminative information from the input data. The highest ratios between the average interclass and intraclass dissimilarities were related to the most accurate classifiers. We can use this information to select a suitable input size to train the classification model. The approach was tested on two data sets related to the fault diagnosis of reciprocating compressors.

Unmanned Aerial Vehicle Operating Mode Classification Using Deep Residual Learning Feature Extraction

Unmanned Aerial Vehicles (UAVs) undoubtedly pose many security challenges. We need only look to the December 2018 Gatwick Airport incident for an example of the disruption UAVs can cause. In total, 1000 flights were grounded for 36 h over the Christmas period which was estimated to cost over 50 million pounds. In this paper, we introduce a novel approach which considers UAV detection as an imagery classification problem. We consider signal representations Power Spectral Density (PSD); Spectrogram, Histogram and raw IQ constellation as graphical images presented to a deep Convolution Neural Network (CNN) ResNet50 for feature extraction. Pre-trained on ImageNet, transfer learning is utilised to mitigate the requirement for a large signal dataset. We evaluate performance through machine learning classifier Logistic Regression. Three popular UAVs are classified in different modes; switched on; hovering; flying; flying with video; and no UAV present, creating a total of 10 classes. Our results, validated with 5-fold cross validation and an independent dataset, show PSD representation to produce over 91% accuracy for 10 classifications. Our paper treats UAV detection as an imagery classification problem by presenting signal representations as images to a ResNet50, utilising the benefits of transfer learning and outperforming previous work in the field.

Signal Representations Via SIP p-frames and SIP Bessel Multipliers in Separable Banach Spaces

Digital signals are often modeled as functions in Banach spaces, such as the ubiquitous [Formula: see text] spaces. The frame theory in Banach spaces induces flexible representations of signals due to the robustness and redundancy of frames. Nevertheless, the lack of inner product in general Banach spaces limits the direct representations of signals in Banach spaces under a given basis or frame. In this paper, we introduce the concept of semi-inner product (SIP) [Formula: see text]-Bessel multipliers to extend the flexibility of signal representations in separable Banach spaces, where [Formula: see text]. These multipliers are defined as composition of analysis operator of an SIP-I Bessel sequence, a multiplication with a fixed sequence and synthesis operator of an SIP-II Bessel sequence. The basic properties of the SIP [Formula: see text]-Bessel multipliers are investigated. Moreover, as special cases, characterizations of [Formula: see text]-Riesz bases related to signal representations are given, and the multipliers for [Formula: see text]-Riesz bases are discussed. We show that SIP [Formula: see text]-Bessel multipliers for [Formula: see text]-Riesz bases are invertible. Finally, the continuity of SIP [Formula: see text]-Bessel multipliers with respect to their parameters is investigated. The results theoretically show that the SIP [Formula: see text]-Bessel multipliers offer a larger range of freedom than frames on signal representations in Banach spaces.

On the generalizability of diffusion MRI signal representations across acquisition parameters, sequences and tissue types: chronicles of the MEMENTO challenge

Diffusion MRI (dMRI) has become an invaluable tool to assess the microstructural organization of brain tissue. Depending on the specific acquisition settings, the dMRI signal encodes specific properties of the underlying diffusion process. In the last two decades, several signal representations have been proposed to fit the dMRI signal and decode such properties. Most methods, however, are tested and developed on a limited amount of data, and their applicability to other acquisition schemes remains unknown. With this work, we aimed to shed light on the generalizability of existing dMRI signal representations to different diffusion encoding parameters and brain tissue types. To this end, we organized a community challenge - named MEMENTO, making available the same datasets for fair comparisons across algorithms and techniques. We considered two state-of-the-art diffusion datasets, including single-diffusion-encoding (SDE) spin-echo data from a human brain with over 3820 unique diffusion weightings (the MASSIVE dataset), and double (oscillating) diffusion encoding data (DDE/DODE) of a mouse brain including over 2520 unique data points. A subset of the data sampled in 5 different voxels was openly distributed, and the challenge participants were asked to predict the remaining part of the data. After one year, eight participant teams submitted a total of 80 signal fits. For each submission, we evaluated the mean squared error, the variance of the prediction error and the Bayesian information criteria. Most predictions predicted either multi-shell SDE data (37%) or DODE data (22%), followed by cartesian SDE data (19%) and DDE (18%). Most submissions predicted the signals measured with SDE remarkably well, with the exception of low and very strong diffusion weightings. The prediction of DDE and DODE data seemed more challenging, likely because none of the submissions explicitly accounted for diffusion time and frequency. Next to the choice of the model, decisions on fit procedure and hyperparameters play a major role in the prediction performance, highlighting the importance of optimizing and reporting such choices. This work is a community effort to highlight strength and limitations of the field at representing dMRI acquired with trending encoding schemes, gaining insights into how different models generalize to different tissue types and fiber configurations over a large range of diffusion encodings.