Regularizing Common Spatial Patterns to Improve BCI Designs: Unified Theory and New Algorithms
read more
Citations
Deep learning with convolutional neural networks for EEG decoding and visualization.
A Review of Classification Algorithms for EEG-based Brain-Computer Interfaces: A 10-year Update
EEGNet: A Compact Convolutional Network for EEG-based Brain-Computer Interfaces
Review of the BCI Competition IV
Learning Temporal Information for Brain-Computer Interface Using Convolutional Neural Networks
References
The Nature of Statistical Learning Theory
The Nature Of Statistical Learning Theory
A review of classification algorithms for EEG-based brain–computer interfaces
A well-conditioned estimator for large-dimensional covariance matrices
Optimal spatial filtering of single trial EEG during imagined hand movement
Related Papers (5)
Frequently Asked Questions (9)
Q2. What are the future works in "Regularizing common spatial patterns to improve bci designs: unified theory and new algorithms" ?
Future work could deal with investigating performances of RCSP algorithms with very small training sets, so as to reduce BCI calibration time, in the line of their previous studies [ 13 ] [ 29 ].
Q3. What is the purpose of the CSP?
CSP aims at learning spatial filters which maximize the variance of band-pass filtered EEG signals from one class while minimizing their variance from the other class [4][3].
Q4. What is the simplest way to solve the problem of a filtered EEG signal?
CSP relying on covariance matrix estimates, such estimates can suffer from noise or small training sets, and thus benefit from regularization.
Q5. What is the purpose of the paper?
for RCSP, the eigenvectors corresponding to the lowest eigenvalues of M1 minimize Eq. 5, and as such maximize the penalty term (which should be minimized).
Q6. What is the purpose of the article?
With CSP, the eigenvectors corresponding to both the largest and smallest eigenvalues of M (see Section II) are used as the spatial filters, as they respectively maximize and minimize Eq. 1 [4].
Q7. What is the objective function of the RCSP?
in order to obtain the filters which maximize C2 while minimizing C1, the authors also need to maximize the following objective function:JP2(w) = wTC2wwTC1w + αP (w) (8)which is achieved by using the eigenvectors corresponding to the largest eigenvalues of M2 = (C1 + αK)−1C2 as the filters w.
Q8. What is the simplest way to solve the problem of rescaling a ?
CSP uses the spatial filters w which extremize the following function:J(w) = wTXT1 X1w wTXT2 X2w = wTC1w wTC2w (1)where T denotes transpose, Xi is the data matrix for class i (with the training samples as rows and the channels as columns) and Ci is the spatial covariance matrix from class i, assuming a zero mean for EEG signals.
Q9. What is the purpose of the method?
More precisely, such a method consists in adding a regularization term to the CSP objective function in order to penalize solutions (i.e., resulting spatial filters) that do not satisfy a given prior.