Evidence of Variabilities in EEG Dynamics During Motor Imagery-Based Multiclass Brain–Computer Interface

Saha, Simanto; Ahmed, Khawza I.; Mostafa, Raqibul; Hadjileontiadis, Leontios; Khandoker, Ahsan H.

doi:10.1109/tnsre.2017.2778178

Cited by 62 publications

(47 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A set of generalized BCI frameworks would be more feasible to implement as compared to a common BCI framework for all users. Because, it is evident to observe significant intersubject variability in EEG signals (Saha et al, 2017b). Successful quantification of inter-subject associativity may suggest clustering of subjects, each cluster having subjects with EEG signal characteristics that are similar or can be interpreted in an intersubject context.…”

Section: The Concept Of Inter-subject Associativitymentioning

confidence: 99%

“…Recent advances in deep learning methods demonstrate a potential application that alleviates intraand inter-subject variability in BCI settings (Chiarelli et al, 2018;Fahimi et al, 2018). Meanwhile, recent studies suggest that the quantification of inter-subject associativity could be equally important to increase the efficacy of exclusively machine learning-based transfer learning strategies for covariate shift adaptation (Kang et al, 2009;Kang and Choi, 2014;Wronkiewicz et al, 2015;Saha et al, 2017bSaha et al, , 2019Perdikis et al, 2018).…”

Section: The Concept Of Inter-subject Associativitymentioning

confidence: 99%

“…Since the motor learning processes differ across individuals (Herzfeld and Shadmehr, 2014;Wu et al, 2014), significant inter-subject variability in motor behavior is anticipated that manifests in the task-specific electrical activities in the cortico-subcortical networks (Seghier and Price, 2018). Consequently, the cortical activity observed in electroencephalogram (EEG) varies across subjects during MI, impeding its utility for BCI applications (Saha et al, 2017b). A study has suggested that time-variant brain functions cause unreliable EEG signatures with poor reproducibility even within a particular subject (Meyer et al, 2013).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Intra- and Inter-subject Variability in EEG-Based Sensorimotor Brain Computer Interface: A Review

Saha

Baumert

2020

Front. Comput. Neurosci.

Self Cite

155

109

View full text Add to dashboard Cite

Brain computer interfaces (BCI) for the rehabilitation of motor impairments exploit sensorimotor rhythms (SMR) in the electroencephalogram (EEG). However, the neurophysiological processes underpinning the SMR often vary over time and across subjects. Inherent intra-and inter-subject variability causes covariate shift in data distributions that impede the transferability of model parameters amongst sessions/subjects. Transfer learning includes machine learning-based methods to compensate for inter-subject and inter-session (intra-subject) variability manifested in EEG-derived feature distributions as a covariate shift for BCI. Besides transfer learning approaches, recent studies have explored psychological and neurophysiological predictors as well as inter-subject associativity assessment, which may augment transfer learning in EEG-based BCI. Here, we highlight the importance of measuring inter-session/subject performance predictors for generalized BCI frameworks for both normal and motor-impaired people, reducing the necessity for tedious and annoying calibration sessions and BCI training.

show abstract

Section: The Concept Of Inter-subject Associativitymentioning

confidence: 99%

Section: The Concept Of Inter-subject Associativitymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Intra- and Inter-subject Variability in EEG-Based Sensorimotor Brain Computer Interface: A Review

Saha

Baumert

2020

Front. Comput. Neurosci.

Self Cite

155

109

View full text Add to dashboard Cite

show abstract

“…Despite several recent advances, most of the MI-based BCI applications are still limited to the laboratory due to their long calibration time. As the literature shows [6]- [8], due to considerable inter-subject and inter-session variations, a reliable machine learning model that performs well across all sessions and subjects has not been feasible yet. Consequently, a 20-30 minutes calibration phase at the beginning of each new session is typically conducted to acquire sufficient labeled data to train the subject-specific BCI model.…”

Section: Introductionmentioning

confidence: 99%

Weighted Transfer Learning for Improving Motor Imagery-Based Brain–Computer Interface

Azab

Mihaylova

Ang

et al. 2019

IEEE Trans. Neural Syst. Rehabil. Eng.

128

View full text Add to dashboard Cite

One of the major limitations of motor imagery (MI)based brain-computer interface (BCI) is its long calibration time. Due to between sessions/subjects variations in the properties of brain signals, typically a large amount of training data needs to be collected at the beginning of each session to calibrate the parameters of the BCI system for the target user. In this paper, we propose a novel transfer learning approach on the classification domain to reduce the calibration time without sacrificing the classification accuracy of MI-BCI. Thus, when only few subject-specific trials are available for training, the estimation of the classification parameters is improved by incorporating previously recorded data from other users. For this purpose, a regularization parameter is added to the objective function of the classifier to make the classification parameters as close as possible to the classification parameters of the previous users who have feature spaces similar to that of the target subject. In this study, a new similarity measure based on the kullback leibler divergence (KL) is used to measure similarity between two feature spaces obtained using subject-specific common spatial patterns (CSP). The proposed transfer learning approach is applied on the logistic regression classifier and evaluated using three datasets. The results showed that compared to the subject-specific classifier, the proposed weighted transfer learning classifier improved the classification results particularly when few subject-specific trials were available for training (p < 0.05). Importantly, this improvement was more pronounced for users with medium and poor accuracy. Moreover, the statistical results showed that the proposed weighted transfer learning classifier performed significantly better than the considered comparable baseline algorithms.

show abstract

“…In this regard, various measures are reported to quantify the complex dynamics of elicited brain activity, like Kolmogorov complexity [ 12 ], Permutation Entropy, Sample Entropy, and its derived modification termed Fuzzy Entropy [ 13 ] that provides a fuzzy boundary for similarity measurements [ 14 ], or even the fusion of Entropy estimators to achieve the complementarity among different features, as developed in [ 15 ]. However, extraction of ERD/S dynamics using Entropy-based pattern estimation is hampered by several factors like movement artifacts during recording, temporal stability of mirroring activation over several sessions differs notably between MI time intervals [ 16 ], low EEG signal-to-noise ratio, poor performance in small-sample setting [ 17 ], and inter-subject variability in EEG Dynamics [ 18 ]. Hence, the reliability of Entropy-based estimators may be limited by several factors like lacking continuity, robustness to noise, and biasing derived from superimposed trends in signals.…”

Section: Introductionmentioning

confidence: 99%

Entropy-Based Estimation of Event-Related De/Synchronization in Motor Imagery Using Vector-Quantized Patterns

Velásquez-Martínez

Caicedo-Acosta

Castellanos-Domínguez

2020

Entropy

View full text Add to dashboard Cite

Assessment of brain dynamics elicited by motor imagery (MI) tasks contributes to clinical and learning applications. In this regard, Event-Related Desynchronization/Synchronization (ERD/S) is computed from Electroencephalographic signals, which show considerable variations in complexity. We present an Entropy-based method, termed VQEnt, for estimation of ERD/S using quantized stochastic patterns as a symbolic space, aiming to improve their discriminability and physiological interpretability. The proposed method builds the probabilistic priors by assessing the Gaussian similarity between the input measured data and their reduced vector-quantized representation. The validating results of a bi-class imagine task database (left and right hand) prove that VQEnt holds symbols that encode several neighboring samples, providing similar or even better accuracy than the other baseline sample-based algorithms of Entropy estimation. Besides, the performed ERD/S time-series are close enough to the trajectories extracted by the variational percentage of EEG signal power and fulfill the physiological MI paradigm. In BCI literate individuals, the VQEnt estimator presents the most accurate outcomes at a lower amount of electrodes placed in the sensorimotor cortex so that reduced channel set directly involved with the MI paradigm is enough to discriminate between tasks, providing an accuracy similar to the performed by the whole electrode set.

show abstract

Evidence of Variabilities in EEG Dynamics During Motor Imagery-Based Multiclass Brain–Computer Interface

Cited by 62 publications

References 62 publications

Intra- and Inter-subject Variability in EEG-Based Sensorimotor Brain Computer Interface: A Review

Intra- and Inter-subject Variability in EEG-Based Sensorimotor Brain Computer Interface: A Review

Weighted Transfer Learning for Improving Motor Imagery-Based Brain–Computer Interface

Entropy-Based Estimation of Event-Related De/Synchronization in Motor Imagery Using Vector-Quantized Patterns

Contact Info

Product

Resources

About