Latent variable method for automatic adaptation to background states in motor imagery BCI
Objective. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states in...
Saved in:
| Published in | Journal of neural engineering Vol. 15; no. 1; pp. 16004 - 16017 |
|---|---|
| Main Authors | , , |
| Format | Journal Article |
| Language | English |
| Published |
England
IOP Publishing
01.02.2018
|
| Subjects | |
| Online Access | Get full text |
Cover
Loading…
| Abstract | Objective. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states into account in an unsupervised way. Approach. We propose a latent variable method that is based on a probabilistic model with a discrete latent variable. In order to estimate the model's parameters, we suggest to use the expectation maximization algorithm. The proposed method is aimed at assessing characteristics of background states without any corresponding data labeling. In the context of asynchronous motor imagery paradigm, we applied this method to the real data from twelve able-bodied subjects with open/closed eyes serving as background states. Main results. We found that the latent variable method improved classification of target states compared to the baseline method (in seven of twelve subjects). In addition, we found that our method was also capable of background states recognition (in six of twelve subjects). Significance. Without any supervised information on background states, the latent variable method provides a way to improve classification in BCI by taking background states into account at the training stage and then by making decisions on target states weighted by posterior probabilities of background states at the prediction stage. |
|---|---|
| AbstractList | Objective. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states into account in an unsupervised way. Approach. We propose a latent variable method that is based on a probabilistic model with a discrete latent variable. In order to estimate the model's parameters, we suggest to use the expectation maximization algorithm. The proposed method is aimed at assessing characteristics of background states without any corresponding data labeling. In the context of asynchronous motor imagery paradigm, we applied this method to the real data from twelve able-bodied subjects with open/closed eyes serving as background states. Main results. We found that the latent variable method improved classification of target states compared to the baseline method (in seven of twelve subjects). In addition, we found that our method was also capable of background states recognition (in six of twelve subjects). Significance. Without any supervised information on background states, the latent variable method provides a way to improve classification in BCI by taking background states into account at the training stage and then by making decisions on target states weighted by posterior probabilities of background states at the prediction stage. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states into account in an unsupervised way.OBJECTIVEBrain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states into account in an unsupervised way.We propose a latent variable method that is based on a probabilistic model with a discrete latent variable. In order to estimate the model's parameters, we suggest to use the expectation maximization algorithm. The proposed method is aimed at assessing characteristics of background states without any corresponding data labeling. In the context of asynchronous motor imagery paradigm, we applied this method to the real data from twelve able-bodied subjects with open/closed eyes serving as background states.APPROACHWe propose a latent variable method that is based on a probabilistic model with a discrete latent variable. In order to estimate the model's parameters, we suggest to use the expectation maximization algorithm. The proposed method is aimed at assessing characteristics of background states without any corresponding data labeling. In the context of asynchronous motor imagery paradigm, we applied this method to the real data from twelve able-bodied subjects with open/closed eyes serving as background states.We found that the latent variable method improved classification of target states compared to the baseline method (in seven of twelve subjects). In addition, we found that our method was also capable of background states recognition (in six of twelve subjects).MAIN RESULTSWe found that the latent variable method improved classification of target states compared to the baseline method (in seven of twelve subjects). In addition, we found that our method was also capable of background states recognition (in six of twelve subjects).Without any supervised information on background states, the latent variable method provides a way to improve classification in BCI by taking background states into account at the training stage and then by making decisions on target states weighted by posterior probabilities of background states at the prediction stage.SIGNIFICANCEWithout any supervised information on background states, the latent variable method provides a way to improve classification in BCI by taking background states into account at the training stage and then by making decisions on target states weighted by posterior probabilities of background states at the prediction stage. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these states is available even during the training stage. Thus there is a need in a method which is capable of taking background states into account in an unsupervised way. We propose a latent variable method that is based on a probabilistic model with a discrete latent variable. In order to estimate the model's parameters, we suggest to use the expectation maximization algorithm. The proposed method is aimed at assessing characteristics of background states without any corresponding data labeling. In the context of asynchronous motor imagery paradigm, we applied this method to the real data from twelve able-bodied subjects with open/closed eyes serving as background states. We found that the latent variable method improved classification of target states compared to the baseline method (in seven of twelve subjects). In addition, we found that our method was also capable of background states recognition (in six of twelve subjects). Without any supervised information on background states, the latent variable method provides a way to improve classification in BCI by taking background states into account at the training stage and then by making decisions on target states weighted by posterior probabilities of background states at the prediction stage. |
| Author | Dagaev, Nikolay Volkova, Ksenia Ossadtchi, Alexei |
| Author_xml | – sequence: 1 givenname: Nikolay surname: Dagaev fullname: Dagaev, Nikolay email: ndagaev@hse.ru organization: National Research University Higher School of Economics Centre for Cognition and Decision Making, Moscow, Russia – sequence: 2 givenname: Ksenia surname: Volkova fullname: Volkova, Ksenia organization: National Research University Higher School of Economics Centre for Cognition and Decision Making, Moscow, Russia – sequence: 3 givenname: Alexei surname: Ossadtchi fullname: Ossadtchi, Alexei organization: National Research University Higher School of Economics Faculty of Computer Science, School of Data Analysis and Artificial Intelligence, Moscow, Russia |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/28718781$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9kD1PwzAQhi0EAlrYmZBHBkrtOHGSESo-KlVigYHJujiXEkjiYjtI_fe4StsBCaY7nd7ndPeMyGFnOiTkgrMbzrJsytOYT6IkiaYAGZPJATndjw73vWQnZOTcB2OCpzk7JidRlvIszfgpeVuAx87Tb7A1FA3SFv27KWllLIXemxZ8rSmUsPKhMx31hhagP5fW9F1JXZiio3VHW-MDUrewRLumd7P5GTmqoHF4vq1j8vpw_zJ7miyeH-ez28VECyH9JCoSjECLlDEds1ynVSwiLGSMUQoxiyGTOZY8LrVkeSVKDrGWpZDAMl4hcjEmV8PelTVfPTqv2tppbBro0PRO8TziXCTh4xC93Eb7osVSrWy4167VTkcIyCGgrXHOYqV0PfztLdSN4kxtvKuNWLWRrAbvAWS_wN3uf5DrAanNSn2Y3nbB0t_xH6aXkfk |
| CODEN | JNEIEZ |
| CitedBy_id | crossref_primary_10_1016_j_bbr_2019_112240 crossref_primary_10_3390_e20010007 crossref_primary_10_3390_s21062173 crossref_primary_10_3389_fnins_2023_1122661 |
| Cites_doi | 10.1016/j.neucom.2016.01.007 10.1109/86.895948 10.1109/TNSRE.2012.2197221 10.1109/10.841330 10.1088/1741-2560/13/3/036018 10.1155/2007/79642 10.1088/1741-2560/10/3/036007 10.1016/S0893-6080(00)00026-5 10.1016/0013-4694(91)90040-B 10.1098/rsta.2012.0222 10.1007/b98874 10.1152/physrev.00027.2016 10.1109/IEMBS.2005.1615701 10.1088/1741-2560/10/4/046003 10.1109/TBME.2005.851521 10.3389/fnhum.2015.00308 10.1016/j.neucom.2015.02.005 10.1371/journal.pone.0080886 10.1088/1741-2560/4/2/R01 10.1016/j.jneumeth.2011.04.037 10.1109/TBME.2004.827062 10.1088/1741-2560/13/6/066022 10.1007/978-0-387-21606-5 10.1109/CNE.2007.369647 10.1109/TPAMI.2010.125 10.1007/s11517-006-0107-4 10.2307/2669645 10.1016/j.neuroimage.2010.03.022 10.1016/j.jneumeth.2015.01.033 10.1111/j.2517-6161.1977.tb01600.x 10.1088/1741-2560/1/2/001 10.1088/1741-2560/3/1/R02 10.1109/MSP.2008.4408441 10.1109/TMAG.2010.2072775 10.1109/7333.948456 10.1007/s11434-008-0547-3 10.1109/TBME.2004.826698 10.1109/86.895947 10.1109/TNSRE.2003.814441 10.1016/S0079-6123(06)59025-9 10.1016/j.clinph.2007.04.027 10.1016/j.jneumeth.2015.08.004 10.1016/j.tins.2007.02.001 10.1109/TBME.2009.2026181 10.1109/TBME.2010.2055564 10.1007/s00521-010-0481-6 10.1007/978-3-642-35289-8_3 10.1016/S0167-8655(01)00075-7 10.1109/TNSRE.2006.875557 |
| ContentType | Journal Article |
| Copyright | 2017 IOP Publishing Ltd |
| Copyright_xml | – notice: 2017 IOP Publishing Ltd |
| DBID | AAYXX CITATION NPM 7X8 |
| DOI | 10.1088/1741-2552/aa8065 |
| DatabaseName | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitle | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic PubMed |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: https://proxy.k.utb.cz/login?url=http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Anatomy & Physiology |
| DocumentTitleAlternate | Latent variable method for automatic adaptation to background states in motor imagery BCI |
| EISSN | 1741-2552 |
| ExternalDocumentID | 28718781 10_1088_1741_2552_aa8065 jneaa8065 |
| Genre | Research Support, Non-U.S. Gov't Journal Article |
| GrantInformation_xml | – fundername: Russian Academic Excellence Project '5-100' |
| GroupedDBID | --- 1JI 4.4 53G 5B3 5GY 5VS 5ZH 7.M 7.Q AAGCD AAJIO AAJKP AALHV AATNI ABHWH ABJNI ABQJV ABVAM ACAFW ACGFS ACHIP AEFHF AENEX AFYNE AKPSB ALMA_UNASSIGNED_HOLDINGS AOAED ASPBG ATQHT AVWKF AZFZN CEBXE CJUJL CRLBU CS3 DU5 EBS EDWGO EJD EMSAF EPQRW EQZZN F5P HAK IHE IJHAN IOP IZVLO KOT LAP M45 N5L N9A NT- NT. P2P PJBAE RIN RO9 ROL RPA SY9 W28 XPP AAYXX ADEQX CITATION 02O 1WK AERVB AHSEE ARNYC BBWZM FEDTE HVGLF JCGBZ NPM Q02 RNS S3P 7X8 |
| ID | FETCH-LOGICAL-c336t-2b5e2ac3700c409c7f432eb64e27a404a869ed14dc609f3d1a4c6d36a081fee13 |
| IEDL.DBID | IOP |
| ISSN | 1741-2560 1741-2552 |
| IngestDate | Thu Jul 10 18:55:33 EDT 2025 Wed Feb 19 02:32:45 EST 2025 Tue Jul 01 01:58:37 EDT 2025 Thu Apr 24 22:58:19 EDT 2025 Wed Aug 21 03:33:55 EDT 2024 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 1 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c336t-2b5e2ac3700c409c7f432eb64e27a404a869ed14dc609f3d1a4c6d36a081fee13 |
| Notes | JNE-101859.R1 ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| PMID | 28718781 |
| PQID | 1921135871 |
| PQPubID | 23479 |
| PageCount | 14 |
| ParticipantIDs | proquest_miscellaneous_1921135871 crossref_citationtrail_10_1088_1741_2552_aa8065 crossref_primary_10_1088_1741_2552_aa8065 pubmed_primary_28718781 iop_journals_10_1088_1741_2552_aa8065 |
| ProviderPackageCode | CITATION AAYXX |
| PublicationCentury | 2000 |
| PublicationDate | 2018-02-01 |
| PublicationDateYYYYMMDD | 2018-02-01 |
| PublicationDate_xml | – month: 02 year: 2018 text: 2018-02-01 day: 01 |
| PublicationDecade | 2010 |
| PublicationPlace | England |
| PublicationPlace_xml | – name: England |
| PublicationTitle | Journal of neural engineering |
| PublicationTitleAbbrev | JNE |
| PublicationTitleAlternate | J. Neural Eng |
| PublicationYear | 2018 |
| Publisher | IOP Publishing |
| Publisher_xml | – name: IOP Publishing |
| References | 44 45 46 47 49 50 51 10 11 55 12 13 14 Myrden A (29) 2016; 13 15 16 Wu W (48) 2006 Perdikis S (53) 2016; 13 17 18 Bishop C M (30) 2006 1 3 4 5 Shenoy P (20) 2006; 3 7 8 9 LaFleur K (52) 2013; 10 21 23 Perdikis S (54) 2015 24 Raza H (22) 2014 Leuthardt E C (2) 2004; 1 26 27 Zander T O (19) 2011; 9 Mühl C (25) 2014; 8 Liyanage S R (28) 2013; 10 32 Lotte F (6) 2007; 4 33 Dempster A P (31) 1977; 39 34 35 36 37 38 39 40 41 42 43 |
| References_xml | – ident: 27 doi: 10.1016/j.neucom.2016.01.007 – ident: 11 doi: 10.1109/86.895948 – start-page: 230 year: 2014 ident: 22 publication-title: IEEE Int. Conf. on Bioinformatics and Biomedicine – volume: 8 start-page: 114 year: 2014 ident: 25 publication-title: Frontiers Neurosci. – ident: 51 doi: 10.1109/TNSRE.2012.2197221 – volume: 9 issn: 1741-2552 year: 2011 ident: 19 publication-title: J. Neural Eng. – ident: 42 doi: 10.1109/10.841330 – volume: 13 issn: 1741-2552 year: 2016 ident: 53 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/13/3/036018 – ident: 4 doi: 10.1155/2007/79642 – year: 2015 ident: 54 – volume: 10 issn: 1741-2552 year: 2013 ident: 28 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/10/3/036007 – ident: 41 doi: 10.1016/S0893-6080(00)00026-5 – ident: 3 doi: 10.1016/0013-4694(91)90040-B – ident: 55 doi: 10.1098/rsta.2012.0222 – ident: 38 doi: 10.1007/b98874 – ident: 1 doi: 10.1152/physrev.00027.2016 – ident: 17 doi: 10.1109/IEMBS.2005.1615701 – volume: 10 issn: 1741-2552 year: 2013 ident: 52 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/10/4/046003 – ident: 43 doi: 10.1109/TBME.2005.851521 – ident: 24 doi: 10.3389/fnhum.2015.00308 – ident: 26 doi: 10.1016/j.neucom.2015.02.005 – ident: 36 doi: 10.1371/journal.pone.0080886 – volume: 4 start-page: R1 issn: 1741-2552 year: 2007 ident: 6 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/4/2/R01 – ident: 23 doi: 10.1016/j.jneumeth.2011.04.037 – ident: 8 doi: 10.1109/TBME.2004.827062 – volume: 13 issn: 1741-2552 year: 2016 ident: 29 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/13/6/066022 – ident: 33 doi: 10.1007/978-0-387-21606-5 – ident: 44 doi: 10.1109/CNE.2007.369647 – ident: 10 doi: 10.1109/TPAMI.2010.125 – ident: 21 doi: 10.1007/s11517-006-0107-4 – ident: 32 doi: 10.2307/2669645 – year: 2006 ident: 30 publication-title: Pattern Recognition and Machinme Learning – ident: 35 doi: 10.1016/j.neuroimage.2010.03.022 – ident: 37 doi: 10.1016/j.jneumeth.2015.01.033 – volume: 39 start-page: 1 issn: 0035-9246 year: 1977 ident: 31 publication-title: J. R. Stat. Soc. doi: 10.1111/j.2517-6161.1977.tb01600.x – volume: 1 start-page: 63 issn: 1741-2552 year: 2004 ident: 2 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/1/2/001 – volume: 3 start-page: R13 issn: 1741-2552 year: 2006 ident: 20 publication-title: J. Neural Eng. doi: 10.1088/1741-2560/3/1/R02 – start-page: 2387 year: 2006 ident: 48 publication-title: 27th Annual Int. Conf. of the Engineering in Medicine and Biology Society – ident: 18 doi: 10.1109/MSP.2008.4408441 – ident: 13 doi: 10.1109/TMAG.2010.2072775 – ident: 15 doi: 10.1109/7333.948456 – ident: 40 doi: 10.1007/s11434-008-0547-3 – ident: 12 doi: 10.1109/TBME.2004.826698 – ident: 16 doi: 10.1109/86.895947 – ident: 7 doi: 10.1109/TNSRE.2003.814441 – ident: 5 doi: 10.1016/S0079-6123(06)59025-9 – ident: 47 doi: 10.1016/j.clinph.2007.04.027 – ident: 46 doi: 10.1016/j.jneumeth.2015.08.004 – ident: 34 doi: 10.1016/j.tins.2007.02.001 – ident: 45 doi: 10.1109/TBME.2009.2026181 – ident: 50 doi: 10.1109/TBME.2010.2055564 – ident: 9 doi: 10.1007/s00521-010-0481-6 – ident: 49 doi: 10.1007/978-3-642-35289-8_3 – ident: 14 doi: 10.1016/S0167-8655(01)00075-7 – ident: 39 doi: 10.1109/TNSRE.2006.875557 |
| SSID | ssj0031790 |
| Score | 2.2089427 |
| Snippet | Objective. Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information... Brain-computer interface (BCI) systems are known to be vulnerable to variabilities in background states of a user. Usually, no detailed information on these... |
| SourceID | proquest pubmed crossref iop |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 16004 |
| SubjectTerms | brain-computer interface electroencephalography latent variable motor imagery probabilistic model |
| Title | Latent variable method for automatic adaptation to background states in motor imagery BCI |
| URI | https://iopscience.iop.org/article/10.1088/1741-2552/aa8065 https://www.ncbi.nlm.nih.gov/pubmed/28718781 https://www.proquest.com/docview/1921135871 |
| Volume | 15 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV1bS91AEB7UvvhiL96OvbCFVvAh5yTZSzb0yUpFRVofFBSEMHsJiDXnoDkF--s7m805oFQpfcvDbDaZnc1-k5n5BuBTaZQvrTRJhkWeCG8xKetCJoV0Ka8Dn7jpEmS_q4MzcXQuzxfgy7wWZjzpP_1DuoxEwVGFfUKcHhGGzhJCwvkIMYQFF-EF10qF9gWHP05mn2EeqKdiNWSQVmkfo_zbHR6cSYs079Nwszt29l_C5eyBY7bJ9XDamqH9_YjL8T_f6BWs9HCU7UbR17DgmzewutuQK35zz7ZZlyDa_XlfhYtjwqVNy36Rex0KrljsPs0I9jKc0oBA_srQ4SSG91k7ZgbtdSgcaRzrSpfu2FXDyDpoyNVNoM-4Z1_3DtfgbP_b6d5B0ndmSCznqk1yI32OlhdpaslBtEUteO6NEj4vUKQCtSq9y4SzKi1r7jIUVjmukABI7X3G12GpGTd-ExgBLofGSmkdCpV6zL1WVmS1oZNTcjWA0WxtKtvTlofuGT-rLnyudRW0VwXtVVF7A9iZj5hEyo5nZD_TolT9vr17Ru7jzCAq2n8hqIKNH09pREkuNJfkdw5gI1rKfNbgjepCZ1v_OMtbWCY8pmNS-DtYam-n_j1hntZ86Gz7D4_q9ps |
| linkProvider | IOP Publishing |
| linkToPdf | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LT9xADLaAShUXSkuBLRSmUovUQ3aTzCPJkdeKbRHlUCR6Sj2PSAjIrkq2Evx6PJksUiuKKnHLYZxJPJ7M59j-DPCx0MoVRuoowSyNhDMYFVUmo0zamFeeT1y3CbIn6uhMfDmX512f07YWZjzpPv19ugxEwUGFXUJcPiAMnUSEhNMBog8LDia2mocXkivuyfNH305nn2Lu6adCRaSXUHEXp3zsLn-cS_M0978hZ3v0DF_Bz9lDh4yTy_600X1z9xef4zPeahmWOljKdsPw1zDn6jewsluTS359y3ZYmyja_oFfgR_HhE_rhv0mN9sXXrHQhZoR_GU4JQFPAsvQ4iSE-VkzZhrNpS8gqS1rS5hu2EXNyEpI5OLa02jcsr390Vs4Gx5-3z-Kug4NkeFcNVGqpUvR8CyODTmKJqsET51WwqUZilhgrgpnE2GNiouK2wSFUZYrJCBSOZfwVViox7VbB0bAy6I2UhqLQsUOU5crI5JK0wlKa9uDwWx9StPRl_suGldlG0bP89JrsPQaLIMGe_D5QWISqDueGPuJFqbs9u_NE-M-zIyipH3ogytYu_GUJApypbkk_7MHa8FaHmb1Xmme5cm7_5xlG16eHgzL49HJ1w1YJIiWhzzxTVhofk3de4JBjd5qTf0eWYj7_w |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Latent+variable+method+for+automatic+adaptation+to+background+states+in+motor+imagery+BCI&rft.jtitle=Journal+of+neural+engineering&rft.au=Dagaev%2C+Nikolay&rft.au=Volkova%2C+Ksenia&rft.au=Ossadtchi%2C+Alexei&rft.date=2018-02-01&rft.eissn=1741-2552&rft.volume=15&rft.issue=1&rft.spage=016004&rft_id=info:doi/10.1088%2F1741-2552%2Faa8065&rft_id=info%3Apmid%2F28718781&rft.externalDocID=28718781 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1741-2560&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1741-2560&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1741-2560&client=summon |