A regularized discriminative framework for EEG analysis with application to brain–computer interface

Decoding working-memory load duringn-back task performance from high channel fNIRS data

Objective.Functional near-infrared spectroscopy (fNIRS) can measure neural activity through blood oxygenation changes in the brain in a wearable form factor, enabling unique applications for research in and outside the lab and in practical occupational settings. fNIRS has proven capable of measuring cognitive states such as mental workload, often using machine learning (ML) based brain-computer interfaces (BCIs). To date, this research has largely relied on probes with channel counts from under ten to several hundred, although recently a new class of wearable NIRS devices featuring thousands of channels has emerged. This poses unique challenges for ML classification, as fNIRS is typically limited by few training trials which results in severely under-determined estimation problems. So far, it is not well understood how such high-resolution data is best leveraged in practical BCIs and whether state-of-the-art or better performance can be achieved.Approach.To address these questions, we propose an ML strategy to classify working-memory load that relies on spatio-temporal regularization and transfer learning from other subjects in a combination that, to our knowledge, has not been used in previous fNIRS BCIs. The approach can be interpreted as an end-to-end generalized linear model and allows for a high degree of interpretability using channel-level or cortical imaging approaches.Main results.We show that using the proposed methodology, it is possible to achieve state-of-the-art decoding performance with high-resolution fNIRS data. We also replicated several state-of-the-art approaches on our dataset of 43 participants wearing a 3198 dual-channel NIRS device while performing then-Back task and show that these existing methodologies struggle in the high-channel regime and are largely outperformed by the proposed pipeline.Significance.Our approach helps establish high-channel NIRS devices as a viable platform for state-of-the-art BCI and opens new applications using this class of headset while also enabling high-resolution model imaging and interpretation.

Adaptive Multimodel Knowledge Transfer Matrix Machine for EEG Classification

The emerging matrix learning methods have achieved promising performances in electroencephalogram (EEG) classification by exploiting the structural information between the columns or rows of feature matrices. Due to the intersubject variability of EEG data, these methods generally need to collect a large amount of labeled individual EEG data, which would cause fatigue and inconvenience to the subjects. Insufficient subject-specific EEG data will weaken the generalization capability of the matrix learning methods in neural pattern decoding. To overcome this dilemma, we propose an adaptive multimodel knowledge transfer matrix machine (AMK-TMM), which can selectively leverage model knowledge from multiple source subjects and capture the structural information of the corresponding EEG feature matrices. Specifically, by incorporating least-squares (LS) loss with spectral elastic net regularization, we first present an LS support matrix machine (LS-SMM) to model the EEG feature matrices. To boost the generalization capability of LS-SMM in scenarios with limited EEG data, we then propose a multimodel adaption method, which can adaptively choose multiple correlated source model knowledge with a leave-one-out cross-validation strategy on the available target training data. We extensively evaluate our method on three independent EEG datasets. Experimental results demonstrate that our method achieves promising performances on EEG classification.

On decoding of rapid motor imagery in a diverse population using a high-density NIRS device

Introduction

Functional near-infrared spectroscopy (fNIRS) aims to infer cognitive states such as the type of movement imagined by a study participant in a given trial using an optical method that can differentiate between oxygenation states of blood in the brain and thereby indirectly between neuronal activity levels. We present findings from an fNIRS study that aimed to test the applicability of a high-density (>3000 channels) NIRS device for use in short-duration (2 s) left/right hand motor imagery decoding in a diverse, but not explicitly balanced, subject population. A side aim was to assess relationships between data quality, self-reported demographic characteristics, and brain-computer interface (BCI) performance, with no subjects rejected from recruitment or analysis.

Methods

BCI performance was quantified using several published methods, including subject-specific and subject-independent approaches, along with a high-density fNIRS decoder previously validated in a separate study.

Results

We found that decoding of motor imagery on this population proved extremely challenging across all tested methods. Overall accuracy of the best-performing method (the high-density decoder) was 59.1 +/- 6.7% after excluding subjects where almost no optode-scalp contact was made over motor cortex and 54.7 +/- 7.6% when all recorded sessions were included. Deeper investigation revealed that signal quality, hemodynamic responses, and BCI performance were all strongly impacted by the hair phenotypical and demographic factors under investigation, with over half of variance in signal quality explained by demographic factors alone.

Discussion

Our results contribute to the literature reporting on challenges in using current-generation NIRS devices on subjects with long, dense, dark, and less pliable hair types along with the resulting potential for bias. Our findings confirm the need for increased focus on these populations, accurate reporting of data rejection choices across subject intake, curation, and final analysis in general, and signal a need for NIRS optode designs better optimized for the general population to facilitate more robust and inclusive research outcomes.

Sparse Bayesian Learning for End-to-End EEG Decoding

Decoding brain activity from non-invasive electroencephalography (EEG) is crucial for brain-computer interfaces (BCIs) and the study of brain disorders. Notably, end-to-end EEG decoding has gained widespread popularity in recent years owing to the remarkable advances in deep learning research. However, many EEG studies suffer from limited sample sizes, making it difficult for existing deep learning models to effectively generalize to highly noisy EEG data. To address this fundamental limitation, this paper proposes a novel end-to-end EEG decoding algorithm that utilizes a low-rank weight matrix to encode both spatio-temporal filters and the classifier, all optimized under a principled sparse Bayesian learning (SBL) framework. Importantly, this SBL framework also enables us to learn hyperparameters that optimally penalize the model in a Bayesian fashion. The proposed decoding algorithm is systematically benchmarked on five motor imagery BCI EEG datasets ( N=192) and an emotion recognition EEG dataset ( N=45), in comparison with several contemporary algorithms, including end-to-end deep-learning-based EEG decoding algorithms. The classification results demonstrate that our algorithm significantly outperforms the competing algorithms while yielding neurophysiologically meaningful spatio-temporal patterns. Our algorithm therefore advances the state-of-the-art by providing a novel EEG-tailored machine learning tool for decoding brain activity.

A spatial-temporal linear feature learning algorithm for P300-based brain-computer interfaces

Speller brain-computer interface (BCI) systems can help neuromuscular disorders patients write their thoughts by using the electroencephalogram (EEG) signals by just focusing on the speller tasks. For practical speller-based BCI systems, the P300 event-related brain potential is measured by using the EEG signal. In this paper, we design a robust machine-learning algorithm for P300 target detection. The novel spatial-temporal linear feature learning (STLFL) algorithm is proposed to extract high-level P300 features. The STLFL method is a modified linear discriminant analysis technique focusing on the spatial-temporal aspects of information extraction. A new P300 detection structure is then proposed based on the combination of the novel STLFL feature extraction and discriminative restricted Boltzmann machine (DRBM) for the classification approach (STLFL + DRBM). The effectiveness of the proposed technique is evaluated using two state-of-the-art P300 BCI datasets. Across the two available databases, we show that in terms of average target recognition accuracy and standard deviation values, the proposed STLFL + DRBM method outperforms traditional methods by 33.5, 78.5, 93.5, and 98.5% for 1, 5, 10, and 15 repetitions, respectively, in BCI competition III datasets II and by 71.3, 100, 100, and 100% for 1, 5, 10, and 15 repetitions, respectively, in BCI competition II datasets II and by 67.5 ± 4, 84.2 ± 2.5, 93.5 ± 1, 96.3 ± 1, and 98.4 ± 0.5% for rapid serial visual presentation (RSVP) based dataset in repetitions 1-5. The method has some advantages over the existing variants including its efficiency, robustness with a small number of training samples, and a high ability to create discriminative features between classes.

A Regional Smoothing Block Sparse Bayesian Learning Method With Temporal Correlation for Channel Selection in P300 Speller

The P300-based brain-computer interfaces (BCIs) enable participants to communicate by decoding the electroencephalography (EEG) signal. Different regions of the brain correspond to various mental activities. Therefore, removing weak task-relevant and noisy channels through channel selection is necessary when decoding a specific type of activity from EEG. It can improve the recognition accuracy and reduce the training time of the subsequent models. This study proposes a novel block sparse Bayesian-based channel selection method for the P300 speller. In this method, we introduce block sparse Bayesian learning (BSBL) into the channel selection of P300 BCI for the first time and propose a regional smoothing BSBL (RSBSBL) by combining the spatial distribution properties of EEG. The RSBSBL can determine the number of channels adaptively. To ensure practicality, we design an automatic selection iteration strategy model to reduce the time cost caused by the inverse operation of the large-size matrix. We verified the proposed method on two public P300 datasets and on our collected datasets. The experimental results show that the proposed method can remove the inferior channels and work with the classifier to obtain high-classification accuracy. Hence, RSBSBL has tremendous potential for channel selection in P300 tasks.

Electrophysiological Decoding of Spatial and Color Processing in Human Prefrontal Cortex

The prefrontal cortex (PFC) plays a pivotal role in goal-directed cognition, yet its representational code remains an open problem with decoding techniques ineffective in disentangling task-relevant variables from PFC. Here we applied regularized linear discriminant analysis to human scalp EEG data and were able to distinguish a mental-rotation task versus a color-perception task with 87% decoding accuracy. Dorsal and ventral areas in lateral PFC provided the dominant features dissociating the two tasks. Our findings show that EEG can reliably decode two independent task states from PFC and emphasize the PFC dorsal/ventral functional specificity in processing the where rotation task versus the what color task.

Learning Common Time-Frequency-Spatial Patterns for Motor Imagery Classification

The common spatial patterns (CSP) algorithm is the most popular spatial filtering method applied to extract electroencephalogram (EEG) features for motor imagery (MI) based brain-computer interface (BCI) systems. The effectiveness of the CSP algorithm depends on optimal selection of the frequency band and time window from the EEG. Many algorithms have been designed to optimize frequency band selection for CSP, while few algorithms seek to optimize the time window. This study proposes a novel framework, termed common time-frequency-spatial patterns (CTFSP), to extract sparse CSP features from multi-band filtered EEG data in multiple time windows. Specifically, the whole MI period is first segmented into multiple subseries using a sliding time window approach. Then, sparse CSP features are extracted from multiple frequency bands in each time window. Finally, multiple support vector machine (SVM) classifiers with the Radial Basis Function (RBF) kernel are trained to identify the MI tasks and the voting result of these classifiers determines the final output of the BCI. This study applies the proposed CTFSP algorithm to three public EEG datasets (BCI competition III dataset IVa, BCI competition III dataset IIIa, and BCI competition IV dataset 1) to validate its effectiveness, compared against several other state-of-the-art methods. The experimental results demonstrate that the proposed algorithm is a promising candidate for improving the performance of MI-BCI systems.

Capsule Network for ERP Detection in Brain-Computer Interface

Event-related potential (ERP) is bioelectrical activity that occurs in the brain in response to specific events or stimuli, reflecting the electrophysiological changes in the brain during cognitive processes. ERP is important in cognitive neuroscience and has been applied to brain-computer interfaces (BCIs). However, because ERP signals collected on the scalp are weak, mixed with spontaneous electroencephalogram (EEG) signals, and their temporal and spatial features are complex, accurate ERP detection is challenging. Compared to traditional neural networks, the capsule network (CapsNet) replaces scalar-output neurons with vector-output capsules, allowing the various input information to be well preserved in the capsules. In this study, we expect to utilize CapsNet to extract the discriminative spatial-temporal features of ERP and encode them in capsules to reduce the loss of valuable information, thereby improving the ERP detection performance for BCI. Therefore, we propose ERP-CapsNet to perform ERP detection in a BCI speller application. The experimental results on BCI Competition datasets and the Akimpech dataset show that ERP-CapsNet achieves better classification performances than do the state-of-the-art techniques. We also use a decoder to investigate the attributes of ERPs encoded in capsules. The results show that ERP-CapsNet relies on the P300 and P100 components to detect ERP. Therefore, ERP-CapsNet not only acts as an outstanding method for ERP detection, but also provides useful insights into the ERP detection mechanism.

An Introductory Tutorial on Brain–Computer Interfaces and Their Applications

The prospect and potentiality of interfacing minds with machines has long captured human imagination. Recent advances in biomedical engineering, computer science, and neuroscience are making brain–computer interfaces a reality, paving the way to restoring and potentially augmenting human physical and mental capabilities. Applications of brain–computer interfaces are being explored in applications as diverse as security, lie detection, alertness monitoring, gaming, education, art, and human cognition augmentation. The present tutorial aims to survey the principal features and challenges of brain–computer interfaces (such as reliable acquisition of brain signals, filtering and processing of the acquired brainwaves, ethical and legal issues related to brain–computer interface (BCI), data privacy, and performance assessment) with special emphasis to biomedical engineering and automation engineering applications. The content of this paper is aimed at students, researchers, and practitioners to glimpse the multifaceted world of brain–computer interfacing.

Simultaneously Spatiospectral Pattern Learning and Contaminated Trial Pruning for Electroencephalography-Based Brain Computer Interface

Electroencephalography (EEG)-based brain computer interfaces (BCIs) translate motor imagery commands into the movements of an external device (e.g., a robotic arm). The automatic design of spectral and spatial filters is a challenging task, as the frequency bands of the spectral filters must be predefined by previously published studies and given that they may be affected during trials by artifacts and improper motor imagery (MI). This study aimed to eliminate the contaminated trials automatically during classifier training, and to simultaneously learn the spectral and spatial patterns without the need for predefined frequency bands. Compared with previous studies that measured the discriminative power of a frequency band based on mutual information, this study determined the difference of the class conditional probability density function between two MI classes. This information was further shared to measure the contamination level of the trial that simplified the computation. A particle-based approximation technique iteratively constructed a filter bank that extracted discriminative features, and simultaneously removed potentially contaminated trials. The particle weight was estimated by an analysis of variance F-test instead of mutual information as commonly used in previous studies. The experimental results of a publicly available dataset revealed that the proposed method outperformed the other BCI in terms of the classification accuracy. Asymmetrical spatial patterns were found on left- versus right-hand MI classifications. The learnt spectral and spatial patterns were consistent with prior neurophysiological knowledge.

An electroencephalographic signature predicts antidepressant response in major depression

Antidepressants are widely prescribed, but their efficacy relative to placebo is modest, in part because the clinical diagnosis of major depression encompasses biologically heterogeneous conditions. Here, we sought to identify a neurobiological signature of response to antidepressant treatment as compared to placebo. We designed a latent-space machine-learning algorithm tailored for resting-state electroencephalography (EEG) and applied it to data from the largest imaging-coupled, placebo-controlled antidepressant study (n = 309). Symptom improvement was robustly predicted in a manner both specific for the antidepressant sertraline (versus placebo) and generalizable across different study sites and EEG equipment. This sertraline-predictive EEG signature generalized to two depression samples, wherein it reflected general antidepressant medication responsivity and related differentially to a repetitive transcranial magnetic stimulation treatment outcome. Furthermore, we found that the sertraline resting-state EEG signature indexed prefrontal neural responsivity, as measured by concurrent transcranial magnetic stimulation and EEG. Our findings advance the neurobiological understanding of antidepressant treatment through an EEG-tailored computational model and provide a clinical avenue for personalized treatment of depression.

Using the General Linear Model to Improve Performance in fNIRS Single Trial Analysis and Classification: A Perspective

Within a decade, single trial analysis of functional Near Infrared Spectroscopy (fNIRS) signals has gained significant momentum, and fNIRS joined the set of modalities frequently used for active and passive Brain Computer Interfaces (BCI). A great variety of methods for feature extraction and classification have been explored using state-of-the-art Machine Learning methods. In contrast, signal preprocessing and cleaning pipelines for fNIRS often follow simple recipes and so far rarely incorporate the available state-of-the-art in adjacent fields. In neuroscience, where fMRI and fNIRS are established neuroimaging tools, evoked hemodynamic brain activity is typically estimated across multiple trials using a General Linear Model (GLM). With the help of the GLM, subject, channel, and task specific evoked hemodynamic responses are estimated, and the evoked brain activity is more robustly separated from systemic physiological interference using independent measures of nuisance regressors, such as short-separation fNIRS measurements. When correctly applied in single trial analysis, e.g., in BCI, this approach can significantly enhance contrast to noise ratio of the brain signal, improve feature separability and ultimately lead to better classification accuracy. In this manuscript, we provide a brief introduction into the GLM and show how to incorporate it into a typical BCI preprocessing pipeline and cross-validation. Using a resting state fNIRS data set augmented with synthetic hemodynamic responses that provide ground truth brain activity, we compare the quality of commonly used fNIRS features for BCI that are extracted from (1) conventionally preprocessed signals, and (2) signals preprocessed with the GLM and physiological nuisance regressors. We show that the GLM-based approach can provide better single trial estimates of brain activity as well as a new feature type, i.e., the weight of the individual and channel-specific hemodynamic response function (HRF) regressor. The improved estimates yield features with higher separability, that significantly enhance accuracy in a binary classification task when compared to conventional preprocessing—on average +7.4% across subjects and feature types. We propose to adapt this well-established approach from neuroscience to the domain of single-trial analysis and preprocessing wherever the classification of evoked brain activity is of concern, for instance in BCI.

Enhance decoding of pre-movement EEG patterns for brain-computer interfaces

OBJECTIVE

In recent years, brain-computer interface (BCI) systems based on electroencephalography (EEG) have developed rapidly. However, the decoding of voluntary finger pre-movements from EEG is still a challenge for BCIs. This study aimed to analyze the pre-movement EEG features in time and frequency domains and design an efficient method to decode the movement-related patterns.

APPROACH

In this study, we first investigated the EEG features induced by the intention of left and right finger movements. Specifically, the movement-related cortical potential (MRCP) and event-related desynchronization (ERD) features were extracted using discriminative canonical pattern matching (DCPM) and common spatial patterns (CSP), respectively. Then, the two types of features were classified by two fisher discriminant analysis (FDA) classifiers, respectively. Their decision values were further assembled to facilitate the classification. To verify the validity of the proposed method, a private dataset containing 12 subjects and a public dataset from BCI competition II were used for estimating the classification accuracy.

MAIN RESULTS

As a result, for the private dataset, the combination of DCPM and CSP achieved an average accuracy of 80.96%, which was 5.08% higher than the single DCPM method (p   <  0.01) and 10.23% higher than the single CSP method (p   <  0.01). Notably, the highest accuracy could achieve 91.5% for the combination method. The test accuracy of dataset IV of BCI competition II was 90%, which was equal to the best result in the existing literature.

SIGNIFICANCE

The results demonstrate the MRCP and ERD features of pre-movements contain significantly discriminative information, which are complementary to each other, and thereby could be well recognized by the proposed combination method of DCPM and CSP. Therefore, this study provides a promising approach for the decoding of pre-movement EEG patterns, which is significant for the development of BCIs.

Information Theoretic Feature Transformation Learning for Brain Interfaces

OBJECTIVE

A variety of pattern analysis techniques for model training in brain interfaces exploit neural feature dimensionality reduction based on feature ranking and selection heuristics. In the light of broad evidence demonstrating the potential sub-optimality of ranking-based feature selection by any criterion, we propose to extend this focus with an information theoretic learning-driven feature transformation concept.

METHODS

We present a maximum mutual information linear transformation and a nonlinear transformation framework derived by a general definition of the feature transformation learning problem. Empirical assessments are performed based on electroencephalographic data recorded during a four class motor imagery brain-computer interface (BCI) task. Exploiting the state-of-the-art methods for initial feature vector construction, we compare the proposed approaches with conventional feature selection-based dimensionality reduction techniques, which are widely used in brain interfaces. Furthermore, for the multi-class problem, we present and exploit a hierarchical graphical model-based BCI decoding system.

RESULTS

Both binary and multi-class decoding analyses demonstrate significantly better performances with the proposed methods.

CONCLUSION

Information theoretic feature transformations are capable of tackling potential confounders of conventional approaches in various settings.

SIGNIFICANCE

We argue that this concept provides significant insights to extend the focus on feature selection heuristics to a broader definition of feature transformation learning in brain interfaces.

A performance based feature selection technique for subject independent MI based BCI

Purpose

Significant research has been conducted in the field of brain computer interface (BCI) algorithm development, however, many of the resulting algorithms are both complex, and specific to a particular user as the most successful methodology can vary between individuals and sessions. The objective of this study was to develop a simple yet effective method of feature selection to improve the accuracy of a subject independent BCI algorithm and streamline the process of BCI algorithm development. Over the past several years, several high precision features have been suggested by researchers to classify different motor imagery tasks. This research applies fourteen of these features as a feature pool that can be used as a reference for future researchers. Additionally, we look for the most efficient feature or feature set with four different classifiers that best differentiates several motor imagery tasks. In this work we have successfully employed a feature fusion method to obtain the best sub-set of features. We have proposed a novel computer aided feature selection method to determine the best set of features for distinguishing between motor imagery tasks in lieu of the manual feature selection that has been performed in past studies. The features selected by this method were then fed into a Linear Discriminant Analysis, K-nearest neighbor, decision tree, or support vector machine classifier for classification to determine the overall performance.

Methods

The methods used were a novel performance based additive feature fusion algorithm working in conjunction with machine learning in order to classify the motor imagery signals into particular states. The data used for this study was collected from BCI competition III dataset IVa.

Result

The result of this algorithm was a classification accuracy of 99% for a subject independent algorithm with less computation cost compared to traditional methods, in addition to multiple feature/classifier combinations that outperform current subject independent methods.

Conclusion

The conclusion of this study and its significance is that it developed a viable methodology for simple, efficient feature selection and BCI algorithm development, which leads to an overall increase in algorithm classification accuracy.

Regularized Group Sparse Discriminant Analysis for P300-Based Brain–Computer Interface

Event-related potentials (ERPs) especially P300 are popular effective features for brain–computer interface (BCI) systems based on electroencephalography (EEG). Traditional ERP-based BCI systems may perform poorly for small training samples, i.e. the undersampling problem. In this study, the ERP classification problem was investigated, in particular, the ERP classification in the high-dimensional setting with the number of features larger than the number of samples was studied. A flexible group sparse discriminative analysis algorithm based on Moreau–Yosida regularization was proposed for alleviating the undersampling problem. An optimization problem with the group sparse criterion was presented, and the optimal solution was proposed by using the regularized optimal scoring method. During the alternating iteration procedure, the feature selection and classification were performed simultaneously. Two P300-based BCI datasets were used to evaluate our proposed new method and compare it with existing standard methods. The experimental results indicated that the features extracted via our proposed method are efficient and provide an overall better P300 classification accuracy compared with several state-of-the-art methods.

Assessment of Visual Pursuit in Patients With Disorders of Consciousness Based on a Brain-Computer Interface

Visual pursuit assessment is extensively applied in the behavioral scale-based clinical examination of patients with disorders of consciousness (DOC). However, this assessment is challenging because it relies on behavioral markers, and these patients severely lack behavioral responses. Brain-computer interfaces (BCIs) may provide a potential solution to detect brain responses to external stimuli without requiring behavioral expressions. A BCI system was designed to simulate visual pursuit detection in the coma recovery scale-revised (CRS-R). The graphical user interface included four buttons, one that moved on the screen and three that did not. These buttons flashed in a random order. The patients were prompted to follow the moving button. Based on the collected electroencephalography data, the algorithm determined whether the patient focused on the moving target. Among the 14 DOC patients who participated in the assessments based on the BCI system and the CRS-R, four patients exhibited visual pursuit, and three were nonresponsive in both assessments. More importantly, seven patients who did not exhibit visual pursuit in CRS-R were detected to be responsive to the moving target stimuli in the BCI assessment. Furthermore, five out of seven recovered consciousness to some degree and showed visual pursuit in the second CRS-R assessment. The proposed BCI system is better able to detect visual pursuit than the behavioral scale-based assessment and thus can assist in clinically evaluating the challenging population of DOC patients.

A Hybrid Network for ERP Detection and Analysis Based on Restricted Boltzmann Machine

Detecting and Please provide the correct one analyzing the event-related potential (ERP) remains an important problem in neuroscience. Due to the low signal-to-noise ratio and complex spatio-temporal patterns of ERP signals, conventional methods usually rely on ensemble averaging technique for reliable detection, which may obliterate subtle but important information in each trial of ERP signals. Inspired by deep learning methods, we propose a novel hybrid network termed ERP-NET. With hybrid deep structure, the proposed network is able to learn complex spatial and temporal patterns from single-trial ERP signals. To verify the effectiveness of ERP-NET, we carried out a few ERP detection experiments that the proposed model achieved cutting-edge performance. The experimental results demonstrate that the patterns learned by the ERP-NET are discriminative ERP components in which the ERP signals are properly characterized. More importantly, as an effective approach to single-trial analysis, ERP-NET is able to discover new ERP patterns which are significant to neuroscience study as well as BCI applications. Therefore, the proposed ERP-NET is a promising tool for the research on ERP signals.

A large scale screening study with a SMR-based BCI: Categorization of BCI users and differences in their SMR activity

Brain-Computer Interfaces (BCIs) are inefficient for a non-negligible part of the population, estimated around 25%. To understand this phenomenon in Sensorimotor Rhythm (SMR) based BCIs, data from a large-scale screening study conducted on 80 novice participants with the Berlin BCI system and its standard machine-learning approach were investigated. Each participant performed one BCI session with resting state Encephalography, Motor Observation, Motor Execution and Motor Imagery recordings and 128 electrodes. A significant portion of the participants (40%) could not achieve BCI control (feedback performance > 70%). Based on the performance of the calibration and feedback runs, BCI users were stratified in three groups. Analyses directed to detect and elucidate the differences in the SMR activity of these groups were performed. Statistics on reactive frequencies, task prevalence and classification results are reported. Based on their SMR activity, also a systematic list of potential reasons leading to performance drops and thus hints for possible improvements of BCI experimental design are given. The categorization of BCI users has several advantages, allowing researchers 1) to select subjects for further analyses as well as for testing new BCI paradigms or algorithms, 2) to adopt a better subject-dependent training strategy and 3) easier comparisons between different studies.

Spatial-Temporal Discriminative Restricted Boltzmann Machine for Event-Related Potential Detection and Analysis

Detecting event-related potential (ERP) is a challenging problem because of its low signal-to-noise ratio and complex spatial-temporal features. Conventional detection methods usually rely on the ensemble averaging technique, which may eliminate subtle but important information in ERP signals and lead to poor detection performance. Inspired by the good performance of discriminative restricted Boltzmann machine (DRBM) in feature extraction and classification, we propose a spatial-temporal DRBM (ST-DRBM) to extract spatial and temporal features for ERP detection. The experimental results and statistical analyses demonstrate that the proposed method is able to achieve state-of-the-art ERP detection performance. The ST-DRBM is not only an effective ERP detector, but also a practical tool for ERP analysis. Based on the proposed model, similar scalp distribution and temporal variations were found in the ERP signals of different sessions, which indicated the feasibility of cross-session ERP detection. Given its state-of-the-art performance and effective analytical technique, ST-DRBM is promising for ERP-based brain-computer interfaces and neuroscience research.

Developing a Three- to Six-State EEG-Based Brain-Computer Interface for a Virtual Robotic Manipulator Control

OBJECTIVE

We develop an electroencephalography (EEG)-based noninvasive brain-computer interface (BCI) system having short training time (15 min) that can be applied for high-performance control of robotic prosthetic systems.

METHODS

A signal processing system for detecting user's mental intent from EEG data based on up to six-state BCI paradigm is developed and used.

RESULTS

We examine the performance of the developed system on experimental data collected from 12 healthy participants and analyzed offline. Out of 12 participants 3 achieve an accuracy of six-state communication in 80%-90% range, while 2 participants do not achieve a satisfactory accuracy. We further implement an online BCI system for control of a virtual 3 degree-of-freedom (dof) prosthetic manipulator and test it with our three best participants. Two participants are able to successfully complete 100% of the test tasks, demonstrating on average the accuracy rate of 80% and requiring 5-10 s to execute a manipulator move. One participant failed to demonstrate a satisfactory performance in online trials.

CONCLUSION

We show that our offline EEG BCI system can correctly identify different motor imageries in EEG data with high accuracy and our online BCI system can be used for control of a virtual 3 dof prosthetic manipulator.

SIGNIFICANCE

Our results prepare foundation for further development of higher performance EEG BCI-based robotic assistive systems and demonstrate that EEG-based BCI may be feasible for robotic control by paralyzed and immobilized individuals.

Across-subject offline decoding of motor imagery from MEG and EEG

Long calibration time hinders the feasibility of brain-computer interfaces (BCI). If other subjects’ data were used for training the classifier, BCI-based neurofeedback practice could start without the initial calibration. Here, we compare methods for inter-subject decoding of left- vs. right-hand motor imagery (MI) from MEG and EEG. Six methods were tested on data involving MEG and EEG measurements of healthy participants. Inter-subject decoders were trained on subjects showing good within-subject accuracy, and tested on all subjects, including poor performers. Three methods were based on Common Spatial Patterns (CSP), and three others on logistic regression with l₁ - or l_2,1 -norm regularization. The decoding accuracy was evaluated using (1) MI and (2) passive movements (PM) for training, separately for MEG and EEG. With MI training, the best accuracies across subjects (mean 70.6% for MEG, 67.7% for EEG) were obtained using multi-task learning (MTL) with logistic regression and l_2,1-norm regularization. MEG yielded slightly better average accuracies than EEG. With PM training, none of the inter-subject methods yielded above chance level (58.7%) accuracy. In conclusion, MTL and training with other subject’s MI is efficient for inter-subject decoding of MI. Passive movements of other subjects are likely suboptimal for training the MI classifiers.

Across-subject offline decoding of motor imagery from MEG and EEG

Long calibration time hinders the feasibility of brain-computer interfaces (BCI). If other subjects' data were used for training the classifier, BCI-based neurofeedback practice could start without the initial calibration. Here, we compare methods for inter-subject decoding of left- vs. right-hand motor imagery (MI) from MEG and EEG. Six methods were tested on data involving MEG and EEG measurements of healthy participants. Inter-subject decoders were trained on subjects showing good within-subject accuracy, and tested on all subjects, including poor performers. Three methods were based on Common Spatial Patterns (CSP), and three others on logistic regression with l₁ - or l_2,1 -norm regularization. The decoding accuracy was evaluated using (1) MI and (2) passive movements (PM) for training, separately for MEG and EEG. With MI training, the best accuracies across subjects (mean 70.6% for MEG, 67.7% for EEG) were obtained using multi-task learning (MTL) with logistic regression and l_2,1-norm regularization. MEG yielded slightly better average accuracies than EEG. With PM training, none of the inter-subject methods yielded above chance level (58.7%) accuracy. In conclusion, MTL and training with other subject's MI is efficient for inter-subject decoding of MI. Passive movements of other subjects are likely suboptimal for training the MI classifiers.

Defining and quantifying users' mental imagery-based BCI skills: a first step

OBJECTIVE

While promising for many applications, electroencephalography (EEG)-based brain-computer interfaces (BCIs) are still scarcely used outside laboratories, due to a poor reliability. It is thus necessary to study and fix this reliability issue. Doing so requires the use of appropriate reliability metrics to quantify both the classification algorithm and the BCI user's performances. So far, classification accuracy (CA) is the typical metric used for both aspects. However, we argue in this paper that CA is a poor metric to study BCI users' skills. Here, we propose a definition and new metrics to quantify such BCI skills for mental imagery (MI) BCIs, independently of any classification algorithm.

APPROACH

We first show in this paper that CA is notably unspecific, discrete, training data and classifier dependent, and as such may not always reflect successful self-modulation of EEG patterns by the user. We then propose a definition of MI-BCI skills that reflects how well the user can self-modulate EEG patterns, and thus how well he could control an MI-BCI. Finally, we propose new performance metrics, classDis, restDist and classStab that specifically measure how distinct and stable the EEG patterns produced by the user are, independently of any classifier.

MAIN RESULTS

By re-analyzing EEG data sets with such new metrics, we indeed confirmed that CA may hide some increase in MI-BCI skills or hide the user inability to self-modulate a given EEG pattern. On the other hand, our new metrics could reveal such skill improvements as well as identify when a mental task performed by a user was no different than rest EEG.

SIGNIFICANCE

Our results showed that when studying MI-BCI users' skills, CA should be used with care, and complemented with metrics such as the new ones proposed. Our results also stressed the need to redefine BCI user training by considering the different BCI subskills and their measures. To promote the complementary use of our new metrics, we provide the Matlab code to compute them for free and open-source.

Scenario Screen: A Dynamic and Context Dependent P300 Stimulator Screen Aimed at Wheelchair Navigation Control

P300 spellers have been widely modified to implement nonspelling tasks. In this work, we propose a "scenario" stimulation screen that is a P300 speller variation to command a wheelchair. Our approach utilized a stimulation screen with an image background (scenario snapshot for a wheelchair) and stimulation markers arranged asymmetrically over relevant landmarks, such as suitable paths, doors, windows, and wall signs. Other scenario stimulation screen features were green/blue stimulation marker color scheme, variable Interstimulus Interval, single marker stimulus mode, and optimized stimulus sequence generator. Eighteen able-bodied subjects participated in the experiment; 78% had no experience in BCI usage. A waveform feature analysis and a Mann-Whitney U test performed over the pairs of target and nontarget coherent averages confirmed that 94% of the subjects elicit P300 (p < .005) on this modified stimulator. Least Absolute Shrinkage and Selection Operator optimization and Linear Discriminant Analysis were utilized for the automatic detection of P300. For evaluation with unseen data, target detection was computed (median sensitivity = 1.00 (0.78-1.00)), together with nontarget discrimination (median specificity = 1.00 (0.98-1.00)). The scenario screen adequately elicits P300 and seems suitable for commanding a wheelchair even when users have no previous experience on the BCI spelling task.

Simultaneous acquisition of EEG and NIRS during cognitive tasks for an open access dataset

We provide an open access multimodal brain-imaging dataset of simultaneous electroencephalography (EEG) and near-infrared spectroscopy (NIRS) recordings. Twenty-six healthy participants performed three cognitive tasks: 1) n-back (0-, 2- and 3-back), 2) discrimination/selection response task (DSR) and 3) word generation (WG) tasks. The data provided includes: 1) measured data, 2) demographic data, and 3) basic analysis results. For n-back (dataset A) and DSR tasks (dataset B), event-related potential (ERP) analysis was performed, and spatiotemporal characteristics and classification results for 'target' versus 'non-target' (dataset A) and symbol 'O' versus symbol 'X' (dataset B) are provided. Time-frequency analysis was performed to show the EEG spectral power to differentiate the task-relevant activations. Spatiotemporal characteristics of hemodynamic responses are also shown. For the WG task (dataset C), the EEG spectral power and spatiotemporal characteristics of hemodynamic responses are analyzed, and the potential merit of hybrid EEG-NIRS BCIs was validated with respect to classification accuracy. We expect that the dataset provided will facilitate performance evaluation and comparison of many neuroimaging analysis techniques.

Evaluating the versatility of EEG models generated from motor imagery tasks: An exploratory investigation on upper-limb elbow-centered motor imagery tasks

Electroencephalography (EEG) has recently been considered for use in rehabilitation of people with motor deficits. EEG data from the motor imagery of different body movements have been used, for instance, as an EEG-based control method to send commands to rehabilitation devices that assist people to perform a variety of different motor tasks. However, it is both time and effort consuming to go through data collection and model training for every rehabilitation task. In this paper, we investigate the possibility of using an EEG model from one type of motor imagery (e.g.: elbow extension and flexion) to classify EEG from other types of motor imagery activities (e.g.: open a drawer). In order to study the problem, we focused on the elbow joint. Specifically, nine kinesthetic motor imagery tasks involving the elbow were investigated in twelve healthy individuals who participated in the study. While results reported that models from goal-oriented motor imagery tasks had higher accuracy than models from the simple joint tasks in intra-task testing (e.g., model from elbow extension and flexion task was tested on EEG data collected from elbow extension and flexion task), models from simple joint tasks had higher accuracies than the others in inter-task testing (e.g., model from elbow extension and flexion task tested on EEG data collected from drawer opening task). Simple single joint motor imagery tasks could, therefore, be considered for training models to potentially reduce the number of repetitive data acquisitions and model training in rehabilitation applications.

Real-time analysis on ensemble SVM scores to reduce P300-Speller intensification time

In most Brain-Computer Interface systems, especially the P300-Speller, there must be a harmonized balance between the accuracy and the spelling time. One major drawback of the classical 36-choice P300-Speller is the slow rate of character elicitation. This paper aims to propose a real-time signal processing method to decrease the spelling time by exploiting the score margins of the ensemble Support Vector Machine classifiers during real-time P300-Speller flashes, rather than just getting the classifiers' highest scores. Our experiments were conducted on the dataset of the BCI Competition III and resulted in a successful character rate of over 96% with just approximately 15 to 20 seconds for each character spelling session. As compared with the fixed 31.5 seconds of the best original approach of the competition, our proposed method significantly reduces the required spelling time by over 30% while maintaining the desired classification accuracy.

A Ternary Brain-Computer Interface Based on Single-Trial Readiness Potentials of Self-initiated Fine Movements: A Diversified Classification Scheme

In recent years, the readiness potential (RP), a type of pre-movement neural activity, has been investigated for asynchronous electroencephalogram (EEG)-based brain-computer interfaces (BCIs). Since the RP is attenuated for involuntary movements, a BCI driven by RP alone could facilitate intentional control amid a plethora of unintentional movements. Previous studies have mainly attempted binary single-trial classification of RP. An RP-based BCI with three or more states would expand the options for functional control. Here, we propose a ternary BCI based on single-trial RPs. This BCI classifies amongst an idle state, a left hand and a right hand self-initiated fine movement. A pipeline of spatio-temporal filtering with per participant parameter optimization was used for feature extraction. The ternary classification was decomposed into binary classifications using a decision-directed acyclic graph (DDAG). For each class pair in the DDAG structure, an ordered diversified classifier system (ODCS-DDAG) was used to select the best among various classification algorithms or to combine the results of different classification algorithms. Using EEG data from 14 participants performing self-initiated left or right key presses, punctuated with rest periods, we compared the performance of ODCS-DDAG to a ternary classifier and four popular multiclass decomposition methods using only a single classification algorithm. ODCS-DDAG had the highest performance (0.769 Cohen's Kappa score) and was significantly better than the ternary classifier and two of the four multiclass decomposition methods. Our work supports further study of RP-based BCI for intuitive asynchronous environmental control or augmentative communication.

An Efficient Framework for EEG Analysis with Application to Hybrid Brain Computer Interfaces Based on Motor Imagery and P300

The hybrid brain computer interface (BCI) based on motor imagery (MI) and P300 has been a preferred strategy aiming to improve the detection performance through combining the features of each. However, current methods used for combining these two modalities optimize them separately, which does not result in optimal performance. Here, we present an efficient framework to optimize them together by concatenating the features of MI and P300 in a block diagonal form. Then a linear classifier under a dual spectral norm regularizer is applied to the combined features. Under this framework, the hybrid features of MI and P300 can be learned, selected, and combined together directly. Experimental results on the data set of hybrid BCI based on MI and P300 are provided to illustrate competitive performance of the proposed method against other conventional methods. This provides an evidence that the method used here contributes to the discrimination performance of the brain state in hybrid BCI.

Decoding of top-down cognitive processing for SSVEP-controlled BMI

We present a fast and accurate non-invasive brain-machine interface (BMI) based on demodulating steady-state visual evoked potentials (SSVEPs) in electroencephalography (EEG). Our study reports an SSVEP-BMI that, for the first time, decodes primarily based on top-down and not bottom-up visual information processing. The experimental setup presents a grid-shaped flickering line array that the participants observe while intentionally attending to a subset of flickering lines representing the shape of a letter. While the flickering pixels stimulate the participant’s visual cortex uniformly with equal probability, the participant’s intention groups the strokes and thus perceives a ‘letter Gestalt’. We observed decoding accuracy of 35.81% (up to 65.83%) with a regularized linear discriminant analysis; on average 2.05-fold, and up to 3.77-fold greater than chance levels in multi-class classification. Compared to the EEG signals, an electrooculogram (EOG) did not significantly contribute to decoding accuracies. Further analysis reveals that the top-down SSVEP paradigm shows the most focalised activation pattern around occipital visual areas; Granger causality analysis consistently revealed prefrontal top-down control over early visual processing. Taken together, the present paradigm provides the first neurophysiological evidence for the top-down SSVEP BMI paradigm, which potentially enables multi-class intentional control of EEG-BMIs without using gaze-shifting.

Optimizing Single-Trial EEG Classification by Stationary Matrix Logistic Regression in Brain-Computer Interface

In addition to the noisy and limited spatial resolution characteristics of the electroencephalography (EEG) signal, the intrinsic nonstationarity in the EEG data makes the single-trial EEG classification an even more challenging problem in brain-computer interface (BCI). Variations of the signal properties within a session often result in deteriorated classification performance. This is mainly attributed to the reason that the routine feature extraction or classification method does not take the changes in the signal into account. Although several extensions to the standard feature extraction method have been proposed to reduce the sensitivity to nonstationarity in data, they optimize different objective functions from that of the subsequent classification model, and thereby, the extracted features may not be optimized for the classification. In this paper, we propose an approach that directly optimizes the classifier's discriminativity and robustness against the within-session nonstationarity of the EEG data through a single optimization paradigm, and show that it can greatly improve the performance, in particular for the subjects who have difficulty in controlling a BCI. Moreover, the experimental results on two benchmark data sets demonstrate that our approach significantly outperforms the compared approaches in reducing classification error rates.

A pipeline of spatio-temporal filtering for predicting the laterality of self-initiated fine movements from single trial readiness potentials

OBJECTIVE

In recent years, the readiness potential (RP), a type of pre-movement neural activity, has been investigated for asynchronous electroencephalogram (EEG)-based brain-computer interfaces (BCIs). Since the RP is attenuated for involuntary movements, a BCI driven by RP alone could facilitate intentional control amid a plethora of unintentional movements. Previous studies have attempted single trial classification of RP via spatial and temporal filtering methods, or by combining the RP with event-related desynchronization. However, RP feature extraction remains challenging due to the slow non-oscillatory nature of the potential, its variability among participants and the inherent noise in EEG signals. Here, we propose a participant-specific, individually optimized pipeline of spatio-temporal filtering (PSTF) to improve RP feature extraction for laterality prediction.

APPROACH

PSTF applies band-pass filtering on RP signals, followed by Fisher criterion spatial filtering to maximize class separation, and finally temporal window averaging for feature dimension reduction. Optimal parameters are simultaneously found by cross-validation for each participant. Using EEG data from 14 participants performing self-initiated left or right key presses as well as two benchmark BCI datasets, we compared the performance of PSTF to two popular methods: common spatial subspace decomposition, and adaptive spatio-temporal filtering.

MAIN RESULTS

On the BCI benchmark data sets, PSTF performed comparably to both existing methods. With the key press EEG data, PSTF extracted more discriminative features, thereby leading to more accurate (74.99% average accuracy) predictions of RP laterality than that achievable with existing methods.

SIGNIFICANCE

Naturalistic and volitional interaction with the world is an important capacity that is lost with traditional system-paced BCIs. We demonstrated a significant improvement in fine movement laterality prediction from RP features alone. Our work supports further study of RP-based BCI for intuitive asynchronous control of the environment, such as augmentative communication or wheelchair navigation.

A Novel Fixed Low-Rank Constrained EEG Spatial Filter Estimation with Application to Movie-Induced Emotion Recognition

This paper proposes a novel fixed low-rank spatial filter estimation for brain computer interface (BCI) systems with an application that recognizes emotions elicited by movies. The proposed approach unifies such tasks as feature extraction, feature selection, and classification, which are often independently tackled in a “bottom-up” manner, under a regularized loss minimization problem. The loss function is explicitly derived from the conventional BCI approach and solves its minimization by optimization with a nonconvex fixed low-rank constraint. For evaluation, an experiment was conducted to induce emotions by movies for dozens of young adult subjects and estimated the emotional states using the proposed method. The advantage of the proposed method is that it combines feature selection, feature extraction, and classification into a monolithic optimization problem with a fixed low-rank regularization, which implicitly estimates optimal spatial filters. The proposed method shows competitive performance against the best CSP-based alternatives.

Orthogonal Connectivity Factorization: Interpretable Decomposition of Variability in Correlation Matrices

In many multivariate time series, the correlation structure is nonstationary, that is, it changes over time. The correlation structure may also change as a function of other cofactors, for example, the identity of the subject in biomedical data. A fundamental approach for the analysis of such data is to estimate the correlation structure (connectivities) separately in short time windows or for different subjects and use existing machine learning methods, such as principal component analysis (PCA), to summarize or visualize the changes in connectivity. However, the visualization of such a straightforward PCA is problematic because the ensuing connectivity patterns are much more complex objects than, say, spatial patterns. Here, we develop a new framework for analyzing variability in connectivities using the PCA approach as the starting point. First, we show how to analyze and visualize the principal components of connectivity matrices by a tailor-made rank-two matrix approximation in which we use the outer product of two orthogonal vectors. This leads to a new kind of transformation of eigenvectors that is particularly suited for this purpose and often enables interpretation of the principal component as connectivity between two groups of variables. Second, we show how to incorporate the orthogonality and the rank-two constraint in the estimation of PCA itself to improve the results. We further provide an interpretation of these methods in terms of estimation of a probabilistic generative model related to blind separation of dependent sources. Experiments on brain imaging data give very promising results.

The LDA beamformer: Optimal estimation of ERP source time series using linear discriminant analysis

We introduce a novel beamforming approach for estimating event-related potential (ERP) source time series based on regularized linear discriminant analysis (LDA). The optimization problems in LDA and linearly-constrained minimum-variance (LCMV) beamformers are formally equivalent. The approaches differ in that, in LCMV beamformers, the spatial patterns are derived from a source model, whereas in an LDA beamformer the spatial patterns are derived directly from the data (i.e., the ERP peak). Using a formal proof and MEG simulations, we show that the LDA beamformer is robust to correlated sources and offers a higher signal-to-noise ratio than the LCMV beamformer and PCA. As an application, we use EEG data from an oddball experiment to show how the LDA beamformer can be harnessed to detect single-trial ERP latencies and estimate connectivity between ERP sources. Concluding, the LDA beamformer optimally reconstructs ERP sources by maximizing the ERP signal-to-noise ratio. Hence, it is a highly suited tool for analyzing ERP source time series, particularly in EEG/MEG studies wherein a source model is not available.

EEG-based usability assessment of 3D shutter glasses

OBJECTIVE

Neurotechnology can contribute to the usability assessment of products by providing objective measures of neural workload and can uncover usability impediments that are not consciously perceived by test persons. In this study, the neural processing effort imposed on the viewer of 3D television by shutter glasses was quantified as a function of shutter frequency. In particular, we sought to determine the critical shutter frequency at which the 'neural flicker' vanishes, such that visual fatigue due to this additional neural effort can be prevented by increasing the frequency of the system.

APPROACH

Twenty-three participants viewed an image through 3D shutter glasses, while multichannel electroencephalogram (EEG) was recorded. In total ten shutter frequencies were employed, selected individually for each participant to cover the range below, at and above the threshold of flicker perception. The source of the neural flicker correlate was extracted using independent component analysis and the flicker impact on the visual cortex was quantified by decoding the state of the shutter from the EEG.

MAIN RESULT

Effects of the shutter glasses were traced in the EEG up to around 67 Hz-about 20 Hz over the flicker perception threshold-and vanished at the subsequent frequency level of 77 Hz.

SIGNIFICANCE

The impact of the shutter glasses on the visual cortex can be detected by neurotechnology even when a flicker is not reported by the participants. Potential impact. Increasing the shutter frequency from the usual 50 Hz or 60 Hz to 77 Hz reduces the risk of visual fatigue and thus improves shutter-glass-based 3D usability.

Separable Common Spatio-Spectral Patterns for Motor Imagery BCI Systems

OBJECTIVE

Feature extraction is one of the most important steps in any brain-computer interface (BCI) system. In particular, spatio-spectral feature extraction for motor-imagery BCIs (MI-BCI) has been the focus of several works in the past decade. This paper proposes a novel method, called separable common spatio-spectral patterns (SCSSP), for extraction of discriminant spatio-spectral EEG features in MI-BCIs.

METHODS

Assuming a binary classification problem, SCSSP uses a heteroscedastic matrix-variate Gaussian model for the multiband EEG rhythms, and seeks the spatio-spectral features whose variance is maximized for one brain task and minimized for the other task. Therefore, SCSSP can be considered as a spatio-spectral generalization of the conventional common spatial patterns (CSP) algorithm.

RESULTS

The experimental results on two-class and multiclass motor-imagery data from publicly available BCI Competition datasets demonstrate that the proposed computationally efficient method competes closely with filter-bank CSP (FBCSP), and can even outperform the FBCSP if enough training data are available. Furthermore, SCSSP provides us with a simple measure for ranking the discriminant power of extracted spatio-spectral features, which is not possible in FBCSP.

CONCLUSION

The matrix-variate Gaussian assumption allows the SCSSP method to jointly process the EEG data in both spatial and spectral domains. As a result, compared to the similar solutions in the literature such as FBCSP, the proposed SCSSP method requires significantly lower computations.

SIGNIFICANCE

The proposed computationally efficient spatio-spectral feature extractor is particularly suitable for applications in which the computational power is limited, such as emerging wearable mobile BCI systems.