Two-layer hidden Markov models for multi-class motor imagery classification

Heung-Il Suk; Seong Whan Lee

doi:10.1109/WBD.2010.16

Two-layer hidden Markov models for multi-class motor imagery classification

Heung-Il Suk, Seong Whan Lee

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

12 Citations (Scopus)

Abstract

Classifiers in a high dimensional space based on the signals of multiple electrodes in EEG-based BCIs suffer from the curse of dimensionality due to the limited training dataset. In order to tackle this problem, we design a framework of two-layer hidden Markov models (HMMs) for probabilistic classification of EEG signals. We first independently model the characteristics of EEG signals embedded in each channel for different motor imagery tasks in the lower-layer, and then represent the holistic task-related dynamic EEG patterns in the upper-layer by considering the relationships among channels. From the experimental results based on the dataset II-a of BCI Competition IV (2008), we demonstrated that our method achieved high session-to-session transfer results and was superior to previous methods.

Original language	English
Title of host publication	Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010
Pages	5-8
Number of pages	4
DOIs	https://doi.org/10.1109/WBD.2010.16
Publication status	Published - 2010 Nov 11
Event	Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with the International Conference on Pattern Recognition, ICPR 2010 - Istanbul, Turkey Duration: 2010 Aug 22 → 2010 Aug 22

Other

Other	Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with the International Conference on Pattern Recognition, ICPR 2010
Country	Turkey
City	Istanbul
Period	10/8/22 → 10/8/22

Fingerprint

Keywords

Brain-Computer Interface (BCI)
Electroencephalography (EGG)
Hidden Markov models (HMMs)
Motor imagery classification

ASJC Scopus subject areas

Computer Vision and Pattern Recognition
Clinical Neurology

Cite this

Suk, H-I., & Lee, S. W. (2010). Two-layer hidden Markov models for multi-class motor imagery classification. In Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010 (pp. 5-8). [5581397] https://doi.org/10.1109/WBD.2010.16

Two-layer hidden Markov models for multi-class motor imagery classification. / Suk, Heung-Il ; Lee, Seong Whan.

Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010. 2010. p. 5-8 5581397.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Suk, H-I & Lee, SW 2010, Two-layer hidden Markov models for multi-class motor imagery classification. in Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010., 5581397, pp. 5-8, Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with the International Conference on Pattern Recognition, ICPR 2010, Istanbul, Turkey, 10/8/22. https://doi.org/10.1109/WBD.2010.16

@inproceedings{54f57459392b44f7805ea65b067ba464,

title = "Two-layer hidden Markov models for multi-class motor imagery classification",

abstract = "Classifiers in a high dimensional space based on the signals of multiple electrodes in EEG-based BCIs suffer from the curse of dimensionality due to the limited training dataset. In order to tackle this problem, we design a framework of two-layer hidden Markov models (HMMs) for probabilistic classification of EEG signals. We first independently model the characteristics of EEG signals embedded in each channel for different motor imagery tasks in the lower-layer, and then represent the holistic task-related dynamic EEG patterns in the upper-layer by considering the relationships among channels. From the experimental results based on the dataset II-a of BCI Competition IV (2008), we demonstrated that our method achieved high session-to-session transfer results and was superior to previous methods.",

keywords = "Brain-Computer Interface (BCI), Electroencephalography (EGG), Hidden Markov models (HMMs), Motor imagery classification",

author = "Heung-Il Suk and Lee, {Seong Whan}",

year = "2010",

month = nov

day = "11",

doi = "10.1109/WBD.2010.16",

language = "English",

isbn = "9780769541334",

pages = "5--8",

booktitle = "Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010",

note = "Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with the International Conference on Pattern Recognition, ICPR 2010 ; Conference date: 22-08-2010 Through 22-08-2010",

}

TY - GEN

T1 - Two-layer hidden Markov models for multi-class motor imagery classification

AU - Suk, Heung-Il

AU - Lee, Seong Whan

PY - 2010/11/11

Y1 - 2010/11/11

N2 - Classifiers in a high dimensional space based on the signals of multiple electrodes in EEG-based BCIs suffer from the curse of dimensionality due to the limited training dataset. In order to tackle this problem, we design a framework of two-layer hidden Markov models (HMMs) for probabilistic classification of EEG signals. We first independently model the characteristics of EEG signals embedded in each channel for different motor imagery tasks in the lower-layer, and then represent the holistic task-related dynamic EEG patterns in the upper-layer by considering the relationships among channels. From the experimental results based on the dataset II-a of BCI Competition IV (2008), we demonstrated that our method achieved high session-to-session transfer results and was superior to previous methods.

AB - Classifiers in a high dimensional space based on the signals of multiple electrodes in EEG-based BCIs suffer from the curse of dimensionality due to the limited training dataset. In order to tackle this problem, we design a framework of two-layer hidden Markov models (HMMs) for probabilistic classification of EEG signals. We first independently model the characteristics of EEG signals embedded in each channel for different motor imagery tasks in the lower-layer, and then represent the holistic task-related dynamic EEG patterns in the upper-layer by considering the relationships among channels. From the experimental results based on the dataset II-a of BCI Competition IV (2008), we demonstrated that our method achieved high session-to-session transfer results and was superior to previous methods.

KW - Brain-Computer Interface (BCI)

KW - Electroencephalography (EGG)

KW - Hidden Markov models (HMMs)

KW - Motor imagery classification

UR - http://www.scopus.com/inward/record.url?scp=78049502454&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78049502454&partnerID=8YFLogxK

U2 - 10.1109/WBD.2010.16

DO - 10.1109/WBD.2010.16

M3 - Conference contribution

AN - SCOPUS:78049502454

SN - 9780769541334

SP - 5

EP - 8

BT - Proceedings - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with theInternational Conference on Pattern Recognition, ICPR 2010

T2 - Workshop on Brain Decoding: Pattern Recognition Challenges in Neuroimaging, WBD 2010 - In Conjunction with the International Conference on Pattern Recognition, ICPR 2010

Y2 - 22 August 2010 through 22 August 2010

ER -

Access to Document

10.1109/WBD.2010.16