Feature extraction from the Hermitian manifold for Brain-Computer Interfaces

Jiachen Xu; Vinay  Jayaram; Bernhard Schölkopf; Moritz Grosse-Wentrup

Feature extraction from the Hermitian manifold for Brain-Computer Interfaces

Jiachen Xu, Vinay Jayaram, Bernhard Schölkopf, Moritz Grosse-Wentrup

Publications: Contribution to conference › Other contribution to conference › Peer Reviewed

Abstract

Riemannian geometry-based methods have shown to be effective in many sorts of Brain-Computer Interface (BCI) applications, but are only capable of measuring the power of the measured signal. This paper proposes a set of novel features derived via the Hilbert transform and applies them to the generalized Riemannian manifold, the Hermitian manifold, to see whether the classiﬁcation accuracy beneﬁts from this treatment. To validate these features, we benchmark them with the Mother of All BCI Benchmarks framework, a recently introduced tool to make BCI methods research more reproducible. The results indicate that in some settings the analytic covariance matrix can improve BCI performance.

Original language	English
Number of pages	4
Publication status	Published - 21 Mar 2019
Externally published	Yes
Event	9th International IEEE EMBS Conference on Neural Engineering - Hilton San Francisco Union Square Hotel, San Francisco, United States Duration: 20 Mar 2019 → 23 Mar 2019 https://neuro.embs.org/2019/

Conference

Conference	9th International IEEE EMBS Conference on Neural Engineering
Country/Territory	United States
City	San Francisco
Period	20/03/19 → 23/03/19
Internet address	https://neuro.embs.org/2019/

Austrian Fields of Science 2012

206001 Biomedical engineering

9th International IEEE EMBS Conference on Neural Engineering
Xu, J. (Participant)
20 Mar 2019 → 23 Mar 2019
Activity: Academic events › Participation in ...

Cite this

@conference{9f507947486043e6a8f6e9cbc8af9589,

title = "Feature extraction from the Hermitian manifold for Brain-Computer Interfaces",

abstract = " Riemannian geometry-based methods have shown to be effective in many sorts of Brain-Computer Interface (BCI) applications, but are only capable of measuring the power of the measured signal. This paper proposes a set of novel features derived via the Hilbert transform and applies them to the generalized Riemannian manifold, the Hermitian manifold, to see whether the classiﬁcation accuracy beneﬁts from this treatment. To validate these features, we benchmark them with the Mother of All BCI Benchmarks framework, a recently introduced tool to make BCI methods research more reproducible. The results indicate that in some settings the analytic covariance matrix can improve BCI performance. ",

author = "Jiachen Xu and Vinay Jayaram and Bernhard Sch{\"o}lkopf and Moritz Grosse-Wentrup",

year = "2019",

month = mar,

day = "21",

language = "English",

note = "9th International IEEE EMBS Conference on Neural Engineering ; Conference date: 20-03-2019 Through 23-03-2019",

url = "https://neuro.embs.org/2019/",

}

TY - CONF

T1 - Feature extraction from the Hermitian manifold for Brain-Computer Interfaces

AU - Xu, Jiachen

AU - Jayaram, Vinay

AU - Schölkopf, Bernhard

AU - Grosse-Wentrup, Moritz

PY - 2019/3/21

Y1 - 2019/3/21

N2 - Riemannian geometry-based methods have shown to be effective in many sorts of Brain-Computer Interface (BCI) applications, but are only capable of measuring the power of the measured signal. This paper proposes a set of novel features derived via the Hilbert transform and applies them to the generalized Riemannian manifold, the Hermitian manifold, to see whether the classiﬁcation accuracy beneﬁts from this treatment. To validate these features, we benchmark them with the Mother of All BCI Benchmarks framework, a recently introduced tool to make BCI methods research more reproducible. The results indicate that in some settings the analytic covariance matrix can improve BCI performance.

AB - Riemannian geometry-based methods have shown to be effective in many sorts of Brain-Computer Interface (BCI) applications, but are only capable of measuring the power of the measured signal. This paper proposes a set of novel features derived via the Hilbert transform and applies them to the generalized Riemannian manifold, the Hermitian manifold, to see whether the classiﬁcation accuracy beneﬁts from this treatment. To validate these features, we benchmark them with the Mother of All BCI Benchmarks framework, a recently introduced tool to make BCI methods research more reproducible. The results indicate that in some settings the analytic covariance matrix can improve BCI performance.

M3 - Other contribution to conference

T2 - 9th International IEEE EMBS Conference on Neural Engineering

Y2 - 20 March 2019 through 23 March 2019

ER -

Feature extraction from the Hermitian manifold for Brain-Computer Interfaces

Abstract

Conference

Austrian Fields of Science 2012

Activities

9th International IEEE EMBS Conference on Neural Engineering

Cite this