Regularized common spatial patterns with subject-to-subject transfer of EEG signals

Discrepancy between inter- and intra-subject variability in EEG-based motor imagery brain-computer interface: Evidence from multiple perspectives

Introduction

Inter- and intra-subject variability are caused by the variability of the psychological and neurophysiological factors over time and across subjects. In the application of in Brain-Computer Interfaces (BCI), the existence of inter- and intra-subject variability reduced the generalization ability of machine learning models seriously, which further limited the use of BCI in real life. Although many transfer learning methods can compensate for the inter- and intra-subject variability to some extent, there is still a lack of clear understanding about the change of feature distribution between the cross-subject and cross-session electroencephalography (EEG) signal.

Methods

To investigate this issue, an online platform for motor-imagery BCI decoding has been built in this work. The EEG signal from both the multi-subject (Exp1) and multi-session (Exp2) experiments has been analyzed from multiple perspectives.

Results

Firstly we found that with the similar variability of classification results, the time-frequency response of the EEG signal within-subject in Exp2 is more consistent than cross-subject results in Exp1. Secondly, the standard deviation of the common spatial pattern (CSP) feature has a significant difference between Exp1 and Exp2. Thirdly, for model training, different strategies for the training sample selection should be applied for the cross-subject and cross-session tasks.

Discussion

All these findings have deepened the understanding of inter- and intra-subject variability. They can also guide practice for the new transfer learning methods development in EEG-based BCI. In addition, these results also proved that BCI inefficiency was not caused by the subject's unable to generate the event-related desynchronization/synchronization (ERD/ERS) signal during the motor imagery.

Empirical comparison of deep learning methods for EEG decoding

Electroencephalography (EEG) is a technique that can be used in non-invasive brain-machine interface (BMI) systems to register brain electrical activity. The EEG signals are non-linear and non-stationary, making the decoding procedure a complex task. Deep learning techniques have been successfully applied in several research fields, often improving the results compared with traditional approaches. Therefore, it is believed that these techniques can also improve the process of decoding brain signals in BMI systems. In this work, we present the implementation of two deep learning-based decoders and we compared the results with other state of art deep learning methods. The first decoder uses long short-term memory (LSTM) recurrent neural network and the second, entitled EEGNet-LSTM, combines a well-known neural decoder based on convolutional neural networks, called EEGNet, with some LSTM layers. The decoders have been tested using data set 2a from BCI Competition IV, and the results showed that the EEGNet-LSTM decoder has been approximately 23% better than the competition-winning decoder. A Wilcoxon t-test showed a significant difference between the two decoders (Z = 2.524, p = 0.012). The LSTM-based decoder has been approximately 9% higher than the best decoder from the same competition. However, there was no significant difference (Z = 1.540, p = 0.123). In order to verify the replication of the EEGNet-LSTM decoder on another data, we performed a test with PhysioNet's Physiobank EEG Motor Movement/Imagery dataset. The EEGNet-LSTM presented a higher performance (0.85 accuracy) than the EEGNet (0.82 accuracy). The results of this work can be important for the development of new research, as well as EEG-based BMI systems, which can benefit from the high precision of neural decoders.

Subject adaptation convolutional neural network for EEG-based motor imagery classification

Objective.Deep transfer learning has been widely used to address the nonstationarity of electroencephalogram (EEG) data during motor imagery (MI) classification. However, previous deep learning approaches suffer from limited classification accuracy because the temporal and spatial features cannot be effectively extracted.Approach.Here, we propose a novel end-to-end deep subject adaptation convolutional neural network (SACNN) to handle the problem of EEG-based MI classification. Our proposed model jointly optimizes three modules, i.e. a feature extractor, a classifier, and a subject adapter. Specifically, the feature extractor simultaneously extracts the temporal and spatial features from the raw EEG data using a parallel multiscale convolution network. In addition, we design a subject adapter to reduce the feature distribution shift between the source and target subjects by using the maximum mean discrepancy. By minimizing the classification loss and the distribution discrepancy, the model is able to extract the temporal-spatial features to the prediction of a new subject.Main results.Extensive experiments are carried out on three EEG-based MI datasets, i.e. brain-computer interface (BCI) competition IV dataset IIb, BCI competition III dataset IVa, and BCI competition IV dataset I, and the average accuracy reaches to 86.42%, 81.71% and 79.35% on the three datasets respectively. Furthermore, the statistical analysis also indicates the significant performance improvement of SACNN.Significance.This paper reveals the importance of the temporal-spatial features on EEG-based MI classification task. Our proposed SACNN model can make fully use of the temporal-spatial information to achieve the purpose.

Supervised and Semisupervised Manifold Embedded Knowledge Transfer in Motor Imagery-Based BCI

A long calibration procedure limits the use in practice for a motor imagery (MI)-based brain-computer interface (BCI) system. To tackle this problem, we consider supervised and semisupervised transfer learning. However, it is a challenge for them to cope with high intersession/subject variability in the MI electroencephalographic (EEG) signals. Based on the framework of unsupervised manifold embedded knowledge transfer (MEKT), we propose a supervised MEKT algorithm (sMEKT) and a semisupervised MEKT algorithm (ssMEKT), respectively. sMEKT only has limited labelled samples from a target subject and abundant labelled samples from multiple source subjects. Compared to sMEKT, ssMEKT adds comparably more unlabelled samples from the target subject. After performing Riemannian alignment (RA) and tangent space mapping (TSM), both sMEKT and ssMEKT execute domain adaptation to shorten the differences among subjects. During domain adaptation, to make use of the available samples, two algorithms preserve the source domain discriminability, and ssMEKT preserves the geometric structure embedded in the labelled and unlabelled target domains. Moreover, to obtain a subject-specific classifier, sMEKT minimizes the joint probability distribution shift between the labelled target and source domains, whereas ssMEKT performs the joint probability distribution shift minimization between the unlabelled target domain and all labelled domains. Experimental results on two publicly available MI datasets demonstrate that our algorithms outperform the six competing algorithms, where the sizes of labelled and unlabelled target domains are variable. Especially for the target subjects with 10 labelled samples and 270/190 unlabelled samples, ssMEKT shows 5.27% and 2.69% increase in average accuracy on the two abovementioned datasets compared to the previous best semisupervised transfer learning algorithm (RA-regularized common spatial patterns-weighted adaptation regularization, RA-RCSP-wAR), respectively. Therefore, our algorithms can effectively reduce the need of labelled samples for the target subject, which is of importance for the MI-based BCI application.

A Domain Adaptation-Based Method for Classification of Motor Imagery EEG

Non-stationarity of EEG signals lead to high variability across sessions, which results in low classification accuracy. To reduce the inter-session variability, an unsupervised domain adaptation method is proposed. Arithmetic mean and covariance are exploited to represent the data distribution. First, overall mean alignment is conducted between the source and target data. Then, the data in the target domain is labeled by a classifier trained with the source data. The per-class mean and covariance of the target data are estimated based on the predicted labels. Next, an alignment from the source domain to the target domain is performed according to the covariance of each class in the target domain. Finally, per-class mean adaptation is required after covariance alignment to remove the shift of data distribution caused by covariance alignment. Two public BCI competition datasets, namely the BCI competition III dataset IVa and the BCI competition IV dataset IIa were used to evaluate the proposed method. On both datasets, the proposed method effectively improved classification accuracy.

Analytic beamformer transformation for transfer learning in motion-onset visual evoked potential decoding

OBJECTIVE

While decoders of EEG-based event-related potentials (ERPs) are routinely tailored to the individual user to maximize performance, developing them on populations for individual usage has proven much more challenging. We propose the analytic beamformer transformation (ABT) to extract phase and/or magnitude information from spatiotemporal ERPs in response to motion-onset stimulation.

APPROACH

We have tested ABT on 52 motion-onset visual evoked potential (mVEP) datasets from 26 healthy subjects and compared the classification accuracy of support vector machine (SVM), spatiotemporal beamformer (stBF) and stepwise linear discriminant analysis (SWLDA) when trained on individual subjects and on a population thereof.

MAIN RESULTS

When using phase- and combined phase/magnitude information extracted by ABT, we show significant improvements in accuracy of population-trained classifiers applied to individual users (p<0.001). We also show that 450 epochs are needed for a correct functioning of ABT, which corresponds to 2 minutes of paradigm stimulation.

SIGNIFICANCE

We have shown that ABT can be used to create population-trained mVEP classifiers using a limited number of epochs. We expect this to pertain to other ERPs or synchronous stimulation paradigms, allowing for a more effective, population-based training of visual BCIs. Finally, as ABT renders recordings across subjects more structurally invariant, it could be used for transfer learning purposes in view of plug-and-play BCI applications.

Selective Cross-Subject Transfer Learning Based on Riemannian Tangent Space for Motor Imagery Brain-Computer Interface

A motor imagery (MI) brain-computer interface (BCI) plays an important role in the neurological rehabilitation training for stroke patients. Electroencephalogram (EEG)-based MI BCI has high temporal resolution, which is convenient for real-time BCI control. Therefore, we focus on EEG-based MI BCI in this paper. The identification of MI EEG signals is always quite challenging. Due to high inter-session/subject variability, each subject should spend long and tedious calibration time in collecting amounts of labeled samples for a subject-specific model. To cope with this problem, we present a supervised selective cross-subject transfer learning (sSCSTL) approach which simultaneously makes use of the labeled samples from target and source subjects based on Riemannian tangent space. Since the covariance matrices representing the multi-channel EEG signals belong to the smooth Riemannian manifold, we perform the Riemannian alignment to make the covariance matrices from different subjects close to each other. Then, all aligned covariance matrices are converted into the Riemannian tangent space features to train a classifier in the Euclidean space. To investigate the role of unlabeled samples, we further propose semi-supervised and unsupervised versions which utilize the total samples and unlabeled samples from target subject, respectively. Sequential forward floating search (SFFS) method is executed for source selection. All our proposed algorithms transfer the labeled samples from most suitable source subjects into the feature space of target subject. Experimental results on two publicly available MI datasets demonstrated that our algorithms outperformed several state-of-the-art algorithms using small number of the labeled samples from target subject, especially for good target subjects.

Decoding subjective emotional arousal from EEG during an immersive virtual reality experience

Immersive virtual reality (VR) enables naturalistic neuroscientific studies while maintaining experimental control, but dynamic and interactive stimuli pose methodological challenges. We here probed the link between emotional arousal, a fundamental property of affective experience, and parieto-occipital alpha power under naturalistic stimulation: 37 young healthy adults completed an immersive VR experience, which included rollercoaster rides, while their EEG was recorded. They then continuously rated their subjective emotional arousal while viewing a replay of their experience. The association between emotional arousal and parieto-occipital alpha power was tested and confirmed by (1) decomposing the continuous EEG signal while maximizing the comodulation between alpha power and arousal ratings and by (2) decoding periods of high and low arousal with discriminative common spatial patterns and a long short-term memory recurrent neural network. We successfully combine EEG and a naturalistic immersive VR experience to extend previous findings on the neurophysiology of emotional arousal towards real-world neuroscience.

Causal decoding of individual cortical excitability states

Brain responsiveness to stimulation fluctuates with rapidly shifting cortical excitability state, as reflected by oscillations in the electroencephalogram (EEG). For example, the amplitude of motor-evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) of motor cortex changes from trial to trial. To date, individual estimation of the cortical processes leading to this excitability fluctuation has not been possible. Here, we propose a data-driven method to derive individually optimized EEG classifiers in healthy humans using a supervised learning approach that relates pre-TMS EEG activity dynamics to MEP amplitude. Our approach enables considering multiple brain regions and frequency bands, without defining them a priori, whose compound phase-pattern information determines the excitability. The individualized classifier leads to an increased classification accuracy of cortical excitability states from 57% to 67% when compared to μ-oscillation phase extracted by standard fixed spatial filters. Results show that, for the used TMS protocol, excitability fluctuates predominantly in the μ-oscillation range, and relevant cortical areas cluster around the stimulated motor cortex, but between subjects there is variability in relevant power spectra, phases, and cortical regions. This novel decoding method allows causal investigation of the cortical excitability state, which is critical also for individualizing therapeutic brain stimulation.

Design a novel BCI for neurorehabilitation using concurrent LFP and EEG features: a case study

Brain-computer interfaces (BCI) that enables people with severe motor disabilities to use their brain signals for direct control of objects have attracted increased interest in rehabilitation. To date, no study has investigated feasibility of the BCI framework incorporating both intracortical and scalp signals. Methods: Concurrent local field potential (LFP) from the hand-knob area and scalp EEG were recorded in a paraplegic patient undergoing a spike-based close-loop neurorehabilitation training. Based upon multimodal spatio-spectral feature extraction and Naive Bayes classification, we developed, for the first time, a novel LFP-EEG-BCI for motor intention decoding. A transfer learning (TL) approach was employed to further improve the feasibility. The performance of the proposed LFP-EEG-BCI for four-class upper-limb motor intention decoding was assessed. Results: Using a decision fusion strategy, we showed that the LFP-EEG-BCI significantly (p <0.05) outperformed single modal BCI (LFP-BCI and EEG-BCI) in terms of decoding accuracy with the best performance achieved using regularized common spatial pattern features. Interrogation of feature characteristics revealed discriminative spatial and spectral patterns, which may lead to new insights for better understanding of brain dynamics during different motor imagery tasks and promote development of efficient decoding algorithms. Moreover, we showed that similar classification performance could be obtained with few training trials, therefore highlighting the efficacy of TL. Conclusion: The present findings demonstrated the superiority of the novel LFP-EEG-BCI in motor intention decoding. Significance: This work introduced a novel LFP-EEG-BCI that may lead to new directions for developing practical neurorehabilitation systems with high detection accuracy and multi-paradigm feasibility in clinical applications.

A Review on Signal Processing Approaches to Reduce Calibration Time in EEG-Based Brain-Computer Interface

In an electroencephalogram- (EEG-) based brain–computer interface (BCI), a subject can directly communicate with an electronic device using his EEG signals in a safe and convenient way. However, the sensitivity to noise/artifact and the non-stationarity of EEG signals result in high inter-subject/session variability. Therefore, each subject usually spends long and tedious calibration time in building a subject-specific classifier. To solve this problem, we review existing signal processing approaches, including transfer learning (TL), semi-supervised learning (SSL), and a combination of TL and SSL. Cross-subject TL can transfer amounts of labeled samples from different source subjects for the target subject. Moreover, Cross-session/task/device TL can reduce the calibration time of the subject for the target session, task, or device by importing the labeled samples from the source sessions, tasks, or devices. SSL simultaneously utilizes the labeled and unlabeled samples from the target subject. The combination of TL and SSL can take advantage of each other. For each kind of signal processing approaches, we introduce their concepts and representative methods. The experimental results show that TL, SSL, and their combination can obtain good classification performance by effectively utilizing the samples available. In the end, we draw a conclusion and point to research directions in the future.

Improving the performance of multisubject motor imagery-based BCIs using twin cascaded softmax CNNs

OBJECTIVE

Motor imagery (MI) EEG signals vary greatly among subjects, so scholarly research on motor imagery-based brain-computer interfaces (BCIs) has mainly focused on single-subject systems or subject-dependent systems. However, the single-subject model is applicable only to the target subject, and the small sample number greatly limits the performance of the model. This paper aims to study a convolutional neural network to achieve an adaptable MI-BCI that is applicable to multiple subjects.

APPROACH

In this paper, a twin cascaded softmax convolutional neural network (TCSCNN) is proposed for multisubject MI-BCIs. The proposed TCSCNN is independent and can be applied to any single-subject MI classification CNN model. First, to reduce the influence of individual differences, subject recognition and MI recognition are accomplished simultaneously. A cascaded softmax structure consisting of two softmax layers, related to subject recognition and MI recognition, is subsequently applied. Second, to improve the MI classification precision, a twin network structure is proposed on the basis of ensemble learning. TCSCNN is built by combining a cascaded softmax structure and twin network structure.

MAIN RESULTS

Experiments were conducted on three popular CNN models (EEGNet and Shallow ConvNet and Deep ConvNet from EEGDecoding) and three public datasets (BCI Competition IV datasets 2a and 2b and the High-Gamma dataset) to verify the performance of the proposed TCSCNN. The results show that compared with the state-of-the-art CNN model, the proposed TCSCNN obviously improves the precision and convergence of multisubject MI recognition.

SIGNIFICANCE

This study provides a promising scheme for multisubject MI-BCI, reflecting the progress made in the development and application of MI-BCIs.

OPTICAL+: a frequency-based deep learning scheme for recognizing brain wave signals

A human-computer interaction (HCI) system can be used to detect different categories of the brain wave signals that can be beneficial for neurorehabilitation, seizure detection and sleep stage classification. Research on developing HCI systems using brain wave signals has progressed a lot over the years. However, real-time implementation, computational complexity and accuracy are still a concern. In this work, we address the problem of selecting the appropriate filtering frequency band while also achieving a good system performance by proposing a frequency-based approach using long short-term memory network (LSTM) for recognizing different brain wave signals. Adaptive filtering using genetic algorithm is incorporated for a hybrid system utilizing common spatial pattern and LSTM network. The proposed method (OPTICAL+) achieved an overall average classification error rate of 30.41% and a kappa coefficient value of 0.398, outperforming the state-of-the-art methods. The proposed OPTICAL+ predictor can be used to develop improved HCI systems that will aid in neurorehabilitation and may also be beneficial for sleep stage classification and seizure detection.

Novel channel selection method based on position priori weighted permutation entropy and binary gravity search algorithm

Brain-computer interface (BCI) system based on motor imagery (MI) usually adopts multichannel Electroencephalograph (EEG) signal recording method. However, EEG signals recorded in multi-channel mode usually contain many redundant and artifact information. Therefore, selecting a few effective channels from whole channels may be a means to improve the performance of MI-based BCI systems. We proposed a channel evaluation parameter called position priori weight-permutation entropy (PPWPE), which include amplitude information and position information of a channel. According to the order of PPWPE values, we initially selected half of the channels with large PPWPE value from all sampling electrode channels. Then, the binary gravitational search algorithm (BGSA) was used in searching a channel combination that will be used in determining an optimal channel combination. The features were extracted by common spatial pattern (CSP) method from the final selected channels, and the classifier was trained by support vector machine. The PPWPE + BGSA + CSP channel selection method is validated on two data sets. Results showed that the PPWPE + BGSA + CSP method obtained better mean classification accuracy (88.0% vs. 57.5% for Data set 1 and 91.1% vs. 79.4% for Data set 2) than All-C + CSP method. The PPWPE + BGSA + CSP method can achieve higher classification in fewer channels selected. This method has great potential to improve the performance of MI-based BCI systems.

A brain-computer interface for the continuous, real-time monitoring of working memory load in real-world environments

We developed a brain-computer interface (BCI) able to continuously monitor working memory (WM) load in real-time (considering the last 2.5 s of brain activity). The BCI is based on biomarkers derived from spectral properties of non-invasive electroencephalography (EEG), subsequently classified by a linear discriminant analysis classifier. The BCI was trained on a visual WM task, tested in a real-time visual WM task, and further validated in a real-time cross task (mental arithmetic). Throughout each trial of the cross task, subjects were given real or sham feedback about their WM load. At the end of the trial, subjects were asked whether the feedback provided was real or sham. The high rate of correct answers provided by the subjects validated not only the global behaviour of the WM-load feedback, but also its real-time dynamics. On average, subjects were able to provide a correct answer 82% of the time, with one subject having 100% accuracy. Possible cognitive and motor confounding factors were disentangled to support the claim that our EEG-based markers correspond indeed to WM.

A Between-Subject fNIRS-BCI Study on Detecting Self-Regulated Intention During Walking

OBJECTIVE

Most BCI (brain-computer interface) studies have focused on detecting motion intention from a resting state. However, the dynamic regulation of two motion states, which usually happens in real life, is rarely studied. Besides, popular within-subject methods also require an extensive and time-consuming learning stage when testing on a new subject. This paper proposed a method to discriminate dynamic gait- adjustment intention with strong adaptability for different subjects.

METHODS

Cerebral hemoglobin signals obtained from 30 subjects were studied to decode gait-adjustment intention. Cerebral hemoglobin information was recorded by using fNIRS (functional near infrared spectroscopy) technology. Mathematical morphology filtering was applied to remove zero drift and EWM (Entropy Weight Method) was used to calculate the average hemoglobin values over Regions of Interest (ROIs). The gradient boosting decision tree (GBDT) was utilized to detect the onset of self-regulated intention. A 2-layer-GA-SVM (Genetic Algorithm-Support Vector Machine) model based on stacking algorithm was further proposed to identify the four types of self-regulated intention (speed increase, speed reduction, step increase, and step reduction).

RESULTS

It was found that GBDT had a good performance to detect the onset intention with an average AUC (Area Under Curve) of 0.894. The 2-layer-GA-SVM model boosted the average ACC (accuracy) of four types of intention from 70.6% to 84.4% ( p = 0.005 ) from the single GA-SVM model. Furthermore, the proposed method passed pseudo-online test with the average results as following: AUC = 0.883, TPR (True Positive Rate) = 97.5%, FPR (False Positive Rate) = 0.11%, and LAY (Detection Latency) = -0.52 ± 2.57 seconds for the recognition of gait-adjustment intention; ACC = 80% for the recognition of adjusted gait.

CONCLUSION

The results indicate that it is feasible to decode dynamic gait-adjustment intentions from a motion state for different subjects based on fNIRS technology. It has a potential to realize the practical application of fNIRS-based brain-computer interface technology in controlling walking-assistive devices.

Characterizing Regularization Techniques for Spatial Filter Optimization in Oscillatory EEG Regression Problems : Guidelines Derived from Simulation and Real-World Data

We report on novel supervised algorithms for single-trial brain state decoding. Their reliability and robustness are essential to efficiently perform neurotechnological applications in closed-loop. When brain activity is assessed by multichannel recordings, spatial filters computed by the source power comodulation (SPoC) algorithm allow identifying oscillatory subspaces. They regress to a known continuous trial-wise variable reflecting, e.g. stimulus characteristics, cognitive processing or behavior. In small dataset scenarios, this supervised method tends to overfit to its training data as the involved recordings via electroencephalogram (EEG), magnetoencephalogram or local field potentials generally provide a low signal-to-noise ratio. To improve upon this, we propose and characterize three types of regularization techniques for SPoC: approaches using Tikhonov regularization (which requires model selection via cross-validation), combinations of Tikhonov regularization and covariance matrix normalization as well as strategies exploiting analytical covariance matrix shrinkage. All proposed techniques were evaluated both in a novel simulation framework and on real-world data. Based on the simulation findings, we saw our expectations fulfilled, that SPoC regularization generally reveals the largest benefit for small training sets and under severe label noise conditions. Relevant for practitioners, we derived operating ranges of regularization hyperparameters for cross-validation based approaches and offer open source code. Evaluating all methods additionally on real-world data, we observed an improved regression performance mainly for datasets from subjects with initially poor performance. With this proof-of-concept paper, we provided a generalizable regularization framework for SPoC which may serve as a starting point for implementing advanced techniques in the future.

A Common Spatial Pattern and Wavelet Packet Decomposition Combined Method for EEG-Based Emotion Recognition

To effectively reduce the day-to-day fluctuations and differences in subjects’ brain electroencephalogram (EEG) signals and improve the accuracy and stability of EEG emotion classification, a new EEG feature extraction method based on common spatial pattern (CSP) and wavelet packet decomposition (WPD) is proposed. For the five-day emotion related EEG data of 12 subjects, the CSP algorithm is firstly used to project the raw EEG data into an optimal subspace to extract the discriminative features by maximizing the Kullback-Leibler (KL) divergences between the two categories of EEG data. Then the WPD algorithm is used to decompose the EEG signals into the related features in time-frequency domain. Finally, four state-of-the-art classifiers including Bagging tree, SVM, linear discriminant analysis and Bayesian linear discriminant analysis are used to make binary emotion classification. The experimental results show that with CSP spatial filtering, the emotion classification on the WPD features extracted with bior3.3 wavelet base gets the best accuracy of 0.862, which is 29.3% higher than that of the power spectral density (PSD) feature without CSP preprocessing, is 23% higher than that of the PSD feature with CSP preprocessing, is 1.9% higher than that of the WPD feature extracted with bior3.3 wavelet base without CSP preprocessing, and is 3.2% higher than that of the WPD feature extracted with the rbio6.8 wavelet base without CSP preprocessing. Our proposed method can effectively reduce the variance and non-stationary of the cross-day EEG signals, extract the emotion related features and improve the accuracy and stability of the cross-day EEG emotion classification. It is valuable for the development of robust emotional brain-computer interface applications.

A cognitive brain-computer interface monitoring sustained attentional variations during a continuous task

We introduce a cognitive brain–computer interface based on a continuous performance task for the monitoring of variations of visual sustained attention, i.e. the self-directed maintenance of cognitive focus in non-arousing conditions while possibly ignoring distractors and avoiding mind wandering. We introduce a visual sustained attention continuous performance task with three levels of task difficulty. Pairwise discrimination of these task difficulties from electroencephalographic features was performed using a leave-one-subject-out cross validation approach. Features were selected using the orthogonal forward regression supervised feature selection method. Cognitive load was best predicted using a combination of prefrontal theta power, broad spatial range gamma power, fronto-central beta power, and fronto-central alpha power. Generalization performance estimates for pairwise classification of task difficulty using these features reached 75% for 5 s epochs, and 85% for 30 s epochs.

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.