Journal of Neural Engineering 

About: Journal of Neural Engineering is an academic journal published by IOP Publishing. The journal publishes majorly in the area(s): Medicine & Computer science. It has an ISSN identifier of 1741-2552. Over the lifetime, 3044 publications have been published receiving 123937 citations.

A review of classification algorithms for EEG-based brain–computer interfaces

Abstract: In this paper we review classification algorithms used to design brain–computer interface (BCI) systems based on electroencephalography (EEG). We briefly present the commonly employed algorithms and describe their critical properties. Based on the literature, we compare them in terms of performance and provide guidelines to choose the suitable classification algorithm(s) for a specific BCI.

A Review of Classification Algorithms for EEG-based Brain-Computer Interfaces: A 10-year Update

Abstract: Objective: Most current Electroencephalography (EEG)-based Brain-Computer Interfaces (BCIs) are based on machine learning algorithms. There is a large diversity of classifier types that are used in this field, as described in our 2007 review paper. Now, approximately 10 years after this review publication, many new algorithms have been developed and tested to classify EEG signals in BCIs. The time is therefore ripe for an updated review of EEG classification algorithms for BCIs. Approach: We surveyed the BCI and machine learning literature from 2007 to 2017 to identify the new classification approaches that have been investigated to design BCIs. We synthesize these studies in order to present such algorithms, to report how they were used for BCIs, what were the outcomes, and to identify their pros and cons. Main results: We found that the recently designed classification algorithms for EEG-based BCIs can be divided into four main categories: adaptive classifiers, matrix and tensor classifiers, transfer learning and deep learning, plus a few other miscellaneous classifiers. Among these, adaptive classifiers were demonstrated to be generally superior to static ones, even with unsupervised adaptation. Transfer learning can also prove useful although the benefits of transfer learning remain unpredictable. Riemannian geometry-based methods have reached state-of-the-art performances on multiple BCI problems and deserve to be explored more thoroughly, along with tensor-based methods. Shrinkage linear discriminant analysis and random forests also appear particularly useful for small training samples settings. On the other hand, deep learning methods have not yet shown convincing improvement over state-of-the-art BCI methods. Significance: This paper provides a comprehensive overview of the modern classification algorithms used in EEG-based BCIs, presents the principles of these Review of Classification Algorithms for EEG-based BCI 2 methods and guidelines on when and how to use them. It also identifies a number of challenges to further advance EEG classification in BCI.

EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces

Abstract: Objective Brain-computer interfaces (BCI) enable direct communication with a computer, using neural activity as the control signal. This neural signal is generally chosen from a variety of well-studied electroencephalogram (EEG) signals. For a given BCI paradigm, feature extractors and classifiers are tailored to the distinct characteristics of its expected EEG control signal, limiting its application to that specific signal. Convolutional neural networks (CNNs), which have been used in computer vision and speech recognition to perform automatic feature extraction and classification, have successfully been applied to EEG-based BCIs; however, they have mainly been applied to single BCI paradigms and thus it remains unclear how these architectures generalize to other paradigms. Here, we ask if we can design a single CNN architecture to accurately classify EEG signals from different BCI paradigms, while simultaneously being as compact as possible. Approach In this work we introduce EEGNet, a compact convolutional neural network for EEG-based BCIs. We introduce the use of depthwise and separable convolutions to construct an EEG-specific model which encapsulates well-known EEG feature extraction concepts for BCI. We compare EEGNet, both for within-subject and cross-subject classification, to current state-of-the-art approaches across four BCI paradigms: P300 visual-evoked potentials, error-related negativity responses (ERN), movement-related cortical potentials (MRCP), and sensory motor rhythms (SMR). Main results We show that EEGNet generalizes across paradigms better than, and achieves comparably high performance to, the reference algorithms when only limited training data is available across all tested paradigms. In addition, we demonstrate three different approaches to visualize the contents of a trained EEGNet model to enable interpretation of the learned features. Significance Our results suggest that EEGNet is robust enough to learn a wide variety of interpretable features over a range of BCI tasks. Our models can be found at: https://github.com/vlawhern/arl-eegmodels.

A brain-computer interface using electrocorticographic signals in humans.

Abstract: Brain–computer interfaces (BCIs) enable users to control devices with electroencephalographic (EEG) activity from the scalp or with single-neuron activity from within the brain Both methods have disadvantages: EEG has limited resolution and requires extensive training, while single-neuron recording entails significant clinical risks and has limited stability We demonstrate here for the first time that electrocorticographic (ECoG) activity recorded from the surface of the brain can enable users to control a one-dimensional computer cursor rapidly and accurately We first identified ECoG signals that were associated with different types of motor and speech imagery Over brief training periods of 3–24 min, four patients then used these signals to master closed-loop control and to achieve success rates of 74–100% in a one-dimensional binary task In additional open-loop experiments, we found that ECoG signals at frequencies up to 180 Hz encoded substantial information about the direction of two-dimensional joystick movements Our results suggest that an ECoG-based BCI could provide for people with severe motor disabilities a non-muscular communication and control option that is more powerful than EEG-based BCIs and is potentially more stable and less traumatic than BCIs that use electrodes penetrating the brain M This article features online multimedia enhancements

An optical neural interface: in vivo control of rodent motor cortex with integrated fiberoptic and optogenetic technology

Abstract: Neural interface technology has made enormous strides in recent years but stimulating electrodes remain incapable of reliably targeting specific cell types (e.g. excitatory or inhibitory neurons) within neural tissue. This obstacle has major scientific and clinical implications. For example, there is intense debate among physicians, neuroengineers and neuroscientists regarding the relevant cell types recruited during deep brain stimulation (DBS); moreover, many debilitating side effects of DBS likely result from lack of cell-type specificity. We describe here a novel optical neural interface technology that will allow neuroengineers to optically address specific cell types in vivo with millisecond temporal precision. Channelrhodopsin-2 (ChR2), an algal light-activated ion channel we developed for use in mammals, can give rise to safe, light-driven stimulation of CNS neurons on a timescale of milliseconds. Because ChR2 is genetically targetable, specific populations of neurons even sparsely embedded within intact circuitry can be stimulated with high temporal precision. Here we report the first in vivo behavioral demonstration of a functional optical neural interface (ONI) in intact animals, involving integrated fiberoptic and optogenetic technology. We developed a solid-state laser diode system that can be pulsed with millisecond precision, outputs 20 mW of power at 473 nm, and is coupled to a lightweight, flexible multimode optical fiber, ∼200 µm in diameter. To capitalize on the unique advantages of this system, we specifically targeted ChR2 to excitatory cells in vivo with the CaMKIIα promoter. Under these conditions, the intensity of light exiting the fiber (∼380 mW mm −2 ) was sufficient to drive excitatory neurons in vivo and control motor cortex function with behavioral output in intact rodents. No exogenous chemical cofactor was needed at any point, a crucial finding for in vivo work in large mammals. Achieving modulation of behavior with optical control of neuronal subtypes may give rise to fundamental network-level insights complementary to what electrode methodologies have taught us, and the emerging optogenetic toolkit may find application across a broad range of neuroscience, neuroengineering and clinical questions.