A supervised recalibration protocol for unbiased BCI

This work presents the findings of a research study which investigated the possibility of automating parameter selection for an EEG-based P300-driven Brain-Computer Interface (BCI) using a Genetic Algorithm (GA). Although this approach showed marked advantages over current practices, the GA approach required lengthy execution times which render it infeasible for online utilisation. The GA method was subsequently replaced by the less execution time-intensive N-fold cross-validation (NFCV) for the meta-optimisation of feature extraction and pre-processing parameters using Fisher's Linear Discriminant Analysis (FLDA). In addition, this work sought to improve BCI classification accuracy using a training data collection and application protocol that the authors refer to as 'Intra-session classifier training and implementation'. Command misclassifications are detrimental to BCI operation since they either result in the execution of the wrong command or they require correction action which is time consuming. The intra-session training protocol has been used in prior instances for BCI development however this work is a novel application of the intra-session protocol to P300-based BCIs. Moreover, there is no formal study that investigates the effect of the intra-session protocol on the classification accuracy of BCIs. This work directly addresses this issue. Intra-session classifier training and implementation using NFCV as a basis for automated parameter proposal yields a classification accuracy of 82.94% compared to 45.44% for the inter-session approach using data-insensitive parameters. The findings are significant as intra-session P300-BCI classifier training can be applied to any P300-based BCI regardless of the existing application platform to obtain improved classification performance.

2016

Clinical Brain-Computer Interface (BCI) systems seek to enable paralyzed individuals to operate devices with their brain activity. Non-invasive systems based on electroen-cephalographic (EEG) signals are popular since they avoid risks associated with invasive procedures, but unfortunately EEG signals are inherently noisy, making effective classifiers challenging to develop. Commonly, new classifiers are benchmarked on signals from healthy subjects executing physical movements, under the assumption that the performance will transfer to clinical cases where only imagined movements are possible. Here, we show in contrast that classifiers trained on signals associated with actual movements perform erratically when applied to signals associated with imagined movements. We suggest that this is because the signals lay in different domains. Then, to exploit the different statistical distributions, we apply a domain adaptation technique, Frustratingly Easy Domain Adaptation (FEDA), improving...

IEEE Transactions on Biomedical Engineering, 2006

J. Neural Eng., 2014

Objective. EEG artifacts of non-neural origin can be separated from neural signals by independent component analysis (ICA). It is unclear (1) how robustly recently proposed artifact classifiers transfer to novel users, novel paradigms or changed electrode setups, and (2) how artifact cleaning by a machine learning classifier impacts the performance of brain-computer interfaces (BCIs). Approach. Addressing (1), the robustness of different strategies with respect to the transfer between paradigms and electrode setups of a recently proposed classifier is investigated on offline data from 35 users and 3 EEG paradigms, which contain 6303 expert-labeled components from two ICA and preprocessing variants. Addressing (2), the effect of artifact removal on single-trial BCI classification is estimated on BCI trials from 101 users and 3 paradigms. Main results. We show that (1) the proposed artifact classifier generalizes to completely different EEG paradigms. To obtain similar results under massively reduced electrode setups, a proposed novel strategy improves artifact classification. Addressing (2), ICA artifact cleaning has little influence on average BCI performance when analyzed by state-of-the-art BCI methods. When slow motor-related features are exploited, performance varies strongly between individuals, as artifacts may obstruct relevant neural activity or are inadvertently used for BCI control. Significance. Robustness of the proposed strategies can be reproduced by EEG practitioners as the method is made available as an EEGLAB plug-in.

2015 7th International IEEE/EMBS Conference on Neural Engineering (NER), 2015

Learning from other subjects and/or sessions led to considerable reduction of calibration time in EEGbased BCIs. However, such learning scheme is not straightforward because of the non-stationary nature of EEG signals. In this paper, we propose an adaptive accuracy-weighted ensemble (AAWE) approach that allows tracking non-stationarity in EEG signals and effectively learning from other subjects. It consists of an ensemble of classifiers, each of which is trained using data recorded from one BCI user. to their accuracy in classifying calibration data of current BCI user. These weights are updated using ensemble decision during feedback phase, when there is no information about true class labels. The effectiveness of our approach is demonstrated through an empirical comparison with other state of the art classifiers combination strategies.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2000

System calibration and user training are essential for operating motor imagery based brain-computer interface (BCI) systems. These steps are often unintuitive and tedious for the user, and do not necessarily lead to a satisfactory level of control. We present an Adaptive BCI framework that provides feedback after only minutes of autocalibration in a two-class BCI setup. During operation, the system recurrently reselects only one out of six predefined logarithmic bandpower features (10-13 and 16-24 Hz from Laplacian derivations over C3, Cz, and C4), specifically, the feature that exhibits maximum discriminability. The system then retrains a linear discriminant analysis classifier on all available data and updates the online paradigm with the new model. Every retraining step is preceded by an online outlier rejection.

IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2006

A Brain-Computer Interface (BCI) is a system that allows its users to control external devices with brain activity. Although the proof-of-concept was given decades ago, the reliable translation of user intent into device control commands is still a major challenge. Success requires the effective interaction of two adaptive controllers: the user's brain, which produces brain activity that encodes intent, and the BCI system, which translates that activity into device control commands. In order to facilitate this interaction, many laboratories are exploring a variety of signal analysis techniques to improve the adaptation of the BCI system to the user. In the literature, many machine learning and pattern classification algorithms have been reported to give impressive results when applied to BCI data in offline analyses. However, it is more difficult to evaluate their relative value for actual online use. BCI data competitions have been organized to provide objective formal evaluations of alternative methods. Prompted by the great interest in the first two BCI Competitions, we organized the third BCI Competition to address several of the most difficult and important analysis problems in BCI research. This article describes the data sets that were provided to the competitors and gives an overview of the results. In a series of accompanying articles, the winning teams describe their methods in detail.

IEEE Transactions on Biomedical Engineering, 2007

A study of different on-line adaptive classifiers, using various feature types is presented. Motor imagery brain computer interface (BCI) experiments were carried out with 18 naive able-bodied subjects. Experiments were done with three two-class, cue-based, electroencephalogram (EEG)-based systems. Two continuously adaptive classifiers were tested: adaptive quadratic and linear discriminant analysis. Three feature types were analyzed, adaptive autoregressive parameters, logarithmic band power estimates and the concatenation of both. Results show that all systems are stable and that the concatenation of features with continuously adaptive linear discriminant analysis classifier is the best choice of all. Also, a comparison of the latter with a discontinuously updated linear discriminant analysis, carried out in on-line experiments with six subjects, showed that on-line adaptation performed significantly better than a discontinuous update. Finally a static subject-specific baseline was also provided and used to compare performance measurements of both types of adaptation.

Frontiers in Neuroergonomics, 2022

Current motor imagery-based brain-computer interface (BCI) systems require a long calibration time at the beginning of each session before they can be used with adequate levels of classification accuracy. In particular, this issue can be a significant burden for long term BCI users. This article proposes a novel transfer learning algorithm, called r-KLwDSA, to reduce the BCI calibration time for long-term users. The proposed r-KLwDSA algorithm aligns the user's EEG data collected in previous sessions to the few EEG trials collected in the current session, using a novel linear alignment method. Thereafter, the aligned EEG trials from the previous sessions and the few EEG trials from the current sessions are fused through a weighting mechanism before they are used for calibrating the BCI model. To validate the proposed algorithm, a large dataset containing the EEG data from 11 stroke patients, each performing 18 BCI sessions, was used. The proposed framework demonstrated a signifi...

2011

Brain computer interfaces (BCIs) translate brain activity into computer commands. To enhance the performance of a BCI, it is necessary to improve the feature extraction techniques being applied to decode the users' intentions. Objective comparison methods are needed to analyze different feature extraction techniques. One possibility is to use the classifier performance as a comparative measure. In this paper, we study the behavior of linear discriminant analysis (LDA) when used to distinguish between electroencephalographic (EEG) signals with and without the presence of event related potentials (ERPs).