Highlights • Data augmentation (DA) is increasingly used with deep learning (DL) on EEG • It enhances decoding accuracy left unexplained by 29% on average on the datasets we review • We analyze which specific DA techniques appear to work best for which EEG tasks • We tested various DA techniques on an open motor-imagery task and compared the accuracy gains to demonstrate the usefulness of DA for DL-based EEG analysis • We propose guidelines for reporting parameters for different DA techniques Abstract-Background Data augmentation (DA) has recently been demonstrated to achieve considerable performance gains for deep learning (DL)-increased accuracy and stability and reduced overfitting. Some electroencephalography (EEG) tasks suffer from low samples-to-features ratio, severely reducing DL effectiveness. DA with DL thus holds transformative promise for EEG processing, possibly like DL revolutionized computer vision, etc.-New method We review trends and approaches to DA for DL in EEG to address: Which DA approaches exist and are common for which EEG tasks? What input features are used? And, what kind of accuracy gain can be expected?-Results DA for DL on EEG begun 5 years ago and is steadily used more. We grouped DA techniques (noise addition, generative adversarial networks, sliding windows, sampling, Fourier transform, recombination of segmentation, and others) and EEG tasks (into seizure detection, sleep stages, motor imagery, mental workload, emotion recognition, motor tasks, and visual tasks). DA efficacy across techniques varied considerably. Noise addition and sliding windows provided the highest accuracy boost; mental workload most benefitted from DA. Sliding window, noise addition, and sampling methods most common for seizure detection, mental workload, and sleep stages, respectively.-Comparing with existing methods Percent of decoding accuracy explained by DA beyond unaugmented accuracy varied between 8% for recombination of segmentation and 36% for noise addition and from 14% for motor imagery to 56% for mental workload-29% on average.-Conclusions DA increasingly used and considerably improved DL decoding accuracy on EEG. Additional publications-if adhering to our reporting guidelines-will facilitate more detailed analysis.
Background Cardiac sympathoexcitation leads to ventricular arrhythmias. Spinal anesthesia modulates sympathetic output and can be cardioprotective. However, its effect on the cardio-spinal reflexes and network interactions in the dorsal horn cardiac afferent neurons and the intermediolateral nucleus sympathetic neurons that regulate sympathetic output is not known. The authors hypothesize that spinal bupivacaine reduces cardiac neuronal firing and network interactions in the dorsal horn–dorsal horn and dorsal horn–intermediolateral nucleus that produce sympathoexcitation during myocardial ischemia, attenuating ventricular arrhythmogenesis. Methods Extracellular neuronal signals from the dorsal horn and intermediolateral nucleus neurons were simultaneously recorded in Yorkshire pigs (n = 9) using a 64-channel high-density penetrating microarray electrode inserted at the T2 spinal cord. Dorsal horn and intermediolateral nucleus neural interactions and known markers of cardiac arrhythmogenesis were evaluated during myocardial ischemia and cardiac load–dependent perturbations with intrathecal bupivacaine. Results Cardiac spinal neurons were identified based on their response to myocardial ischemia and cardiac load–dependent perturbations. Spinal bupivacaine did not change the basal activity of cardiac neurons in the dorsal horn or intermediolateral nucleus. After bupivacaine administration, the percentage of cardiac neurons that increased their activity in response to myocardial ischemia was decreased. Myocardial ischemia and cardiac load–dependent stress increased the short-term interactions between the dorsal horn and dorsal horn (324 to 931 correlated pairs out of 1,189 pairs, P < 0.0001), and dorsal horn and intermediolateral nucleus neurons (11 to 69 correlated pairs out of 1,135 pairs, P < 0.0001). Bupivacaine reduced this network response and augmentation in the interactions between dorsal horn–dorsal horn (931 to 38 correlated pairs out of 1,189 pairs, P < 0.0001) and intermediolateral nucleus–dorsal horn neurons (69 to 1 correlated pairs out of 1,135 pairs, P < 0.0001). Spinal bupivacaine reduced shortening of ventricular activation recovery interval and dispersion of repolarization, with decreased ventricular arrhythmogenesis during acute ischemia. Conclusions Spinal anesthesia reduces network interactions between dorsal horn–dorsal horn and dorsal horn–intermediolateral nucleus cardiac neurons in the spinal cord during myocardial ischemia. Blocking short-term coordination between local afferent–efferent cardiac neurons in the spinal cord contributes to a decrease in cardiac sympathoexcitation and reduction of ventricular arrhythmogenesis. Editor’s Perspective What We Already Know about This Topic What This Article Tells Us That Is New
Objective. Motor-imagery (MI) classification base on electroencephalography (EEG) has been long studied in neuroscience and more recently widely used in healthcare applications such as mobile assistive robots and neurorehabilitation. In particular, EEG-based MI classification methods that rely on convolutional neural networks (CNNs) have achieved relatively high classification accuracy. However, naively training CNNs to classify raw EEG data from all channels, especially for high-density EEG, is computationally demanding and requires huge training sets. It often also introduces many irrelevant input features, making it difficult for the CNN to extract the informative ones. This problem is compounded by a dearth of training data, which is particularly acute for MI tasks, because these are cognitively demanding and thus fatigue inducing. Approach. To address these issues, we proposed an end-to-end CNN-based neural network with attentional mechanism together with different data augmentation (DA) techniques. We tested it on two benchmark MI datasets, brain–computer interface (BCI) competition IV 2a and 2b. In addition, we collected a new dataset, recorded using high-density EEG, and containing both MI and motor execution (ME) tasks, which we share with the community. Main results. Our proposed neural-network architecture outperformed all state-of-the-art methods that we found in the literature, with and without DA, reaching an average classification accuracy of 93.6% and 87.83% on BCI 2a and 2b, respectively. We also directly compare decoding of MI and ME tasks. Focusing on MI classification, we find optimal channel configurations and the best DA techniques as well as investigate combining data across participants and the role of transfer learning. Significance. Our proposed approach improves the classification accuracy for MI in the benchmark datasets. In addition, collecting our own dataset enables us to compare MI and ME and investigate various aspects of EEG decoding critical for neuroscience and BCI.
Human urges, desires, and intentions manifest themselves in voluntary action. The final stages of such voluntary action are the muscle contractions that bring it about. Electromyography (EMG) signals measure such muscle contractions. Decoding action contents from EMG require advanced methods for detection, decomposition, processing, and classification and remains a challenge in neuroscience. This study presents a new, time-domain method of classifying EMG for grasping different types of objects. Our proposed method can classify objects with different weights with an accuracy of up to 90%. This progress in neuroscience affects other fields like physiology, brain-computer interfaces, robotics, and so on.
Electromyography (EMG) is a simple, non-invasive, and cost-effective technology for sensing muscle activity. However, EMG is also noisy, complex, and high-dimensional. It has nevertheless been widely used in a host of human-machine-interface applications (electrical wheelchairs, virtual computer mice, prosthesis, robotic fingers, etc.) and in particular to measure reaching and grasping motions of the human hand. Here, we developd a more automated pipeline to predict object weight in a reach-and-grasp task from an open dataset relying only on EMG data. In that we shifted the focus from manual feature-engineering to automated feature-extraction by using raw (filtered) EMG signals and thus letting the algorithms select the features. We further compared intrinsic EMG features, derived from several dimensionality-reduction methods, and then ran some classification algorithms on these low-dimensional representations. We found that the Laplacian Eigenmap algorithm generally outperformed other dimensionality-reduction methods. What is more, optimal classification accuracy was achieved using a combination of Laplacian Eigenmaps (simple-minded) and k-Nearest Neighbors (88% for 3-way classification). Our results, using EMG alone, are comparable to others in the literature that used EMG and EEG together. They also demonstrate the usefulness of dimensionality reduction when classifying movement based on EMG signals and more generally the usefulness of EMG for movement classification.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.