ECoG-based Brain-computer Interface Research Articles

Multi-way feature selection for ECoG-based Brain-Computer Interface

The paper addresses the problem of designing Brain-Computer Interfaces. It investigates feature selection methods in regression, applied to ECoG-based motion decoding. The problem is to predict hand trajectories from the voltage time series of cortical activity. A special characteristic of this problem is the inherently multi-way structure of feature description. The feature description resides in spatial-spectra-temporal domain and includes the voltage time series and their spectral representation. Since electrocorticographic data is highly correlated in temporal, spectral and spatial domains, redundancy of the feature space as well as its dimensionality become a major obstacle for robust solution of the regression problem both in multi-way and flat cases. Feature selection reduces dimensionality and increases model robustness. It plays the crucial role in obtaining adequate predictions.The main contribution of this paper is the following. We propose a filtering feature selection for multi-way data. The proposed method extends quadratic programming feature selection (QPFS) approach. QPFS selects a subset of features by solving a quadratic problem. It incorporates estimates of similarity between features and their relevance to the regression problem. QPFS offers an effective way to leverage similarities between features and their importance. Our modification allows to apply QPFS to multi-way data. By taking the multi-way structure of features into account, the proposed modification reduces computational costs of optimization problem in QPFS. Experimental outcomes demonstrate that the proposed modification improves prediction quality of resultant models. The proposed method is model-free and provides interpretable results, which makes it relevant for knowledge extraction and domain analysis.

Remapping cortical modulation for electrocorticographic brain–computer interfaces: a somatotopy-based approach in individuals with upper-limb paralysis

Objective. Brain–computer interface (BCI) technology aims to provide individuals with paralysis a means to restore function. Electrocorticography (ECoG) uses disc electrodes placed on either the surface of the dura or the cortex to record field potential activity. ECoG has been proposed as a viable neural recording modality for BCI systems, potentially providing stable, long-term recordings of cortical activity with high spatial and temporal resolution. Previously we have demonstrated that a subject with spinal cord injury (SCI) could control an ECoG-based BCI system with up to three degrees of freedom (Wang et al 2013 PLoS One). Here, we expand upon these findings by including brain-control results from two additional subjects with upper-limb paralysis due to amyotrophic lateral sclerosis and brachial plexus injury, and investigate the potential of motor and somatosensory cortical areas to enable BCI control. Approach. Individuals were implanted with high-density ECoG electrode grids over sensorimotor cortical areas for less than 30 d. Subjects were trained to control a BCI by employing a somatotopic control strategy where high-gamma activity from attempted arm and hand movements drove the velocity of a cursor. Main results. Participants were capable of generating robust cortical modulation that was differentiable across attempted arm and hand movements of their paralyzed limb. Furthermore, all subjects were capable of voluntarily modulating this activity to control movement of a computer cursor with up to three degrees of freedom using the somatotopic control strategy. Additionally, for those subjects with electrode coverage of somatosensory cortex, we found that somatosensory cortex was capable of supporting ECoG-based BCI control. Significance. These results demonstrate the feasibility of ECoG-based BCI systems for individuals with paralysis as well as highlight some of the key challenges that must be overcome before such systems are translated to the clinical realm. ClinicalTrials.gov Identifier: NCT01393444.

Optimizing the Detection of Wakeful and Sleep-Like States for Future Electrocorticographic Brain Computer Interface Applications.

Previous studies suggest stable and robust control of a brain-computer interface (BCI) can be achieved using electrocorticography (ECoG). Translation of this technology from the laboratory to the real world requires additional methods that allow users operate their ECoG-based BCI autonomously. In such an environment, users must be able to perform all tasks currently performed by the experimenter, including manually switching the BCI system on/off. Although a simple task, it can be challenging for target users (e.g., individuals with tetraplegia) due to severe motor disability. In this study, we present an automated and practical strategy to switch a BCI system on or off based on the cognitive state of the user. Using a logistic regression, we built probabilistic models that utilized sub-dural ECoG signals from humans to estimate in pseudo real-time whether a person is awake or in a sleep-like state, and subsequently, whether to turn a BCI system on or off. Furthermore, we constrained these models to identify the optimal anatomical and spectral parameters for delineating states. Other methods exist to differentiate wake and sleep states using ECoG, but none account for practical requirements of BCI application, such as minimizing the size of an ECoG implant and predicting states in real time. Our results demonstrate that, across 4 individuals, wakeful and sleep-like states can be classified with over 80% accuracy (up to 92%) in pseudo real-time using high gamma (70–110 Hz) band limited power from only 5 electrodes (platinum discs with a diameter of 2.3 mm) located above the precentral and posterior superior temporal gyrus.

Motor-related brain activity during action observation: a neural substrate for electrocorticographic brain-computer interfaces after spinal cord injury

After spinal cord injury (SCI), motor commands from the brain are unable to reach peripheral nerves and muscles below the level of the lesion. Action observation (AO), in which a person observes someone else performing an action, has been used to augment traditional rehabilitation paradigms. Similarly, AO can be used to derive the relationship between brain activity and movement kinematics for a motor-based brain-computer interface (BCI) even when the user cannot generate overt movements. BCIs use brain signals to control external devices to replace functions that have been lost due to SCI or other motor impairment. Previous studies have reported congruent motor cortical activity during observed and overt movements using magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI). Recent single-unit studies using intracortical microelectrodes also demonstrated that a large number of motor cortical neurons had similar firing rate patterns between overt and observed movements. Given the increasing interest in electrocorticography (ECoG)-based BCIs, our goal was to identify whether action observation-related cortical activity could be recorded using ECoG during grasping tasks. Specifically, we aimed to identify congruent neural activity during observed and executed movements in both the sensorimotor rhythm (10–40 Hz) and the high-gamma band (65–115 Hz) which contains significant movement-related information. We observed significant motor-related high-gamma band activity during AO in both able-bodied individuals and one participant with a complete C4 SCI. Furthermore, in able-bodied participants, both the low and high frequency bands demonstrated congruent activity between action execution and observation. Our results suggest that AO could be an effective and critical procedure for deriving the mapping from ECoG signals to intended movement for an ECoG-based BCI system for individuals with paralysis.

ECoG Classification Research Based on Wavelet Variance and Probabilistic Neural Network

For a typical ECoG-based brain-computer interface system that the subjects task is to imagine movements of either the left small finger or the tongue, a feature extraction algorithm using wavelet variance was proposed. Firstly the wavelet transform was discussed, and the definition and significance of wavelet variance were bring out and taken as feature, 6 channels with most distinctive features were selected from 64 channels for analysis; consequently the EEG data were decomposed using db4 wavelet, the wavelet coefficients variances containing Mu rhythm and Beta rhythm were taken out as features based on ERD/ERS phenomenon, and the features were classified by probabilistic neural network under a optimal spread with an algorithm of cross validation. The result of off-line showed high average classification accuracies of 89.21% and 88% for training and test data were achieved, the wavelet variance has characteristics of more simple and effective and it is suitable for feature extraction in BCI research.

Decoding Finger Flexion from Band-Specific ECoG Signals in Humans

This article presents the method that won the brain-computer interface (BCI) competition IV addressed to the prediction of the finger flexion from electrocorticogram (ECoG) signals. ECoG-based BCIs have recently drawn the attention from the community. Indeed, ECoG can provide higher spatial resolution and better signal quality than classical EEG recordings. It is also more suitable for long-term use. These characteristics allow to decode precise brain activities and to realize efficient ECoG-based neuroprostheses. Signal processing is a very important task in BCIs research for translating brain signals into commands. Here, we present a linear regression method based on the amplitude modulation of band-specific ECoG including a short-term memory for individual finger flexion prediction. The effectiveness of the method was proven by achieving the highest value of correlation coefficient between the predicted and recorded finger flexion values on data set 4 during the BCI competition IV.

Channel Selection and Feature Extraction of ECoG-Based Brain-Computer Interface Using Band Power

Electrocorticography (ECoG) signals have been proved to be associated with different types of motor imagery and have used in brain-computer interface (BCI) research. This paper studies the channel selection and feature extraction using band powers (BP) for a typical ECoG-based BCI system. The subject images movement of left finger or tongue. Firstly, BP features were used for channel selection, and 11 channels which had distinctive features were selected from 64 channels. Then, the features of ECoG signals were extracted using BP, and the dimension of feature vector was reduced with principal components analysis (PCA). Finally, Fisher linear discriminant analysis (LDA) was used for classification. The results of the experiment showed that this algorithm has got good classification accuracy for the test data set.

Robust, long-term control of an electrocorticographic brain-computer interface with fixed parameters

All previous multiple-day brain-computer interface (BCI) experiments have dynamically adjusted the parameterization between the signals measured from the brain and the features used to control the interface. The authors present the results of a multiple-day electrocorticographic (ECoG) BCI experiment. A patient with a subdural electrode array implanted for seizure localization performed tongue motor tasks. After an initial screening and feature selection on the 1st day, 5 consecutive days of cursor-based feedback were performed with a fixed parameterization. Control of the interface was robust throughout all days, with performance increasing to a stable state in which high-frequency ECoG signal could immediately be translated into cursor control. These findings demonstrate that ECoG-based BCIs can be implemented for multiple-day control without the necessity for sophisticated retraining and adaptation.

Classification of imaginary movements in ECoG with a hybrid approach based on multi-dimensional Hilbert-SVM solution

The study presented in this paper shows that electrocorticographic (ECoG) signals can be classified for making use of a human brain–computer interface (BCI) field. The results show that certain invariant phase transition features can be reliably used to classify two types of imagined movements accurately. Those are the left small-finger and tongue movements. Our approach consists of two main parts: channel selection based on Tsallis entropy in Hilbert domain and the nonlinear classification of motor imagery with support vector machines (SVMs). The new approach, based on Hilbert and statistical/entropy measurements, were combined with SVMs based on admissible kernels for classification purposes. The classification accuracy rates were 95% (264/278) and 73% (73/100) for training and testing sets, respectively. The results support the use of classification methods for ECoG-based BCIs.

Two-dimensional movement control using electrocorticographic signals in humans

We show here that a brain–computer interface (BCI) using electrocorticographic activity (ECoG) and imagined or overt motor tasks enables humans to control a computer cursor in two dimensions. Over a brief training period of 12–36 min, each of five human subjects acquired substantial control of particular ECoG features recorded from several locations over the same hemisphere, and achieved average success rates of 53–73% in a two-dimensional four-target center-out task in which chance accuracy was 25%. Our results support the expectation that ECoG-based BCIs can combine high performance with technical and clinical practicality, and also indicate promising directions for further research.

Generalized Features for Electrocorticographic BCIs

This paper studies classifiability of electrocorticographic signals (ECoG) for use in a human brain-computer interface (BCI). The results show that certain spectral features can be reliably used across several subjects to accurately classify different types of movements. Sparse and nonsparse versions of the support vector machine and regularized linear discriminant analysis linear classifiers are assessed and contrasted for the classification problem. In conjunction with a careful choice of features, the classification process automatically and consistently identifies neurophysiological areas known to be involved in the movements. An average two-class classification accuracy of 95% for real movement and around 80% for imagined movement is shown. The high accuracy and generalizability of these results, obtained with as few as 30 data samples per class, support the use of classification methods for ECoG-based BCIs.