MulPi: A Multi-class and Patient-Independent Epileptic Seizure Classifier With Co-Designed Input-stationary Computing-in-SRAM

Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures req...

Celý popis

Uložené v:

Podrobná bibliografia
Vydané v:	IEEE transactions on biomedical circuits and systems Ročník 19; číslo 4; s. 756 - 766
Hlavní autori:	Kim, Bokyung, Huang, Qijia, Taylor, Brady, Zheng, Qilin, Ku, Jonathan, Chen, Yiran, Li, Hai
Médium:	Journal Article
Jazyk:	English
Vydavateľské údaje:	United States IEEE 01.08.2025 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Predmet:	Accuracy Algorithms Artificial neural networks Biological system modeling Brain modeling Chips (memory devices) Classification Classification algorithms Co-design Computation Computer architecture computing-in-memory Convulsions & seizures cross-patient diagnosis automation EEG Electroencephalography Epilepsy Epilepsy - classification Epilepsy - diagnosis Fast Fourier transforms Hardware Humans input-stationary dataflow Multiclass seizure Neural networks Neural Networks, Computer neural processor patient-independence Quality of life Quantization (signal) seizure classification seizure detection seizure prediction Seizures Seizures - classification Seizures - diagnosis Signal Processing, Computer-Assisted SRAM-CIM Static random access memory Training
ISSN:	1932-4545, 1940-9990, 1940-9990
On-line prístup:	Získať plný text
Tagy:	Pridať tag Žiadne tagy, Buďte prvý, kto otaguje tento záznam!

Abstract	Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class ( Mul ) and 2) patient-independent ( Pi ) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN , with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computing-in-memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM : a) input-stationary (IS) CIM for resource-saving, and b) row-wise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm<inline-formula><tex-math notation="LaTeX">{}^{2}</tex-math></inline-formula>.
AbstractList	Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class ( Mul ) and 2) patient-independent ( Pi ) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN , with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computing-in-memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM : a) input-stationary (IS) CIM for resource-saving, and b) row-wise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm<inline-formula><tex-math notation="LaTeX">{}^{2}</tex-math></inline-formula>. Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class (Mul) and 2) patient-independent (Pi) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN, with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computing-in-memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM: a) input-stationary (IS) CIM for resource-saving, and b) row-wise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm[Formula Omitted]. Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class (Mul) and 2) patient-independent (Pi) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN, with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computingin- memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM: a) input-stationary (IS) CIM for resource-saving, and b) rowwise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm2.Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class (Mul) and 2) patient-independent (Pi) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN, with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computingin- memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM: a) input-stationary (IS) CIM for resource-saving, and b) rowwise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm2. Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives to epilepsy patients, delivering them from fatal danger and increasing the quality of life. Authentic detection and prediction of seizures require 1) multi-class (Mul) and 2) patient-independent (Pi) classification. Recent implementable chips for seizure classification rarely satisfy the two requirements due to restricted resources in small chips; therefore, high efficiency is imperative along with accuracy. This paper introduces an efficient MulPi chip, fabricated for the first time to simultaneously fulfill multi-class and patient independence, based on a co-design approach. We develop a 5-layer convolutional neural network (CNN), MulPiCNN, with advanced training techniques for lightness and accuracy. At the hardware level, our SRAM-based chip leverages computing-in-memory (CIM) for efficiency. The fabricated MulPi chip is distinguished from prior CIMs in two folds, namely ISRW-CIM: a) input-stationary (IS) CIM for resource-saving, and b) row-wise (RW) computing to address a challenge of SRAM CIM, empowered by our novel 2T-Hadamard product unit (HPU). MulPi outperforms state-of-the-art chips with 98.5% sensitivity and 99.2% specificity, classifying in 0.12s and 0.348mm^{2}$.
Author	Taylor, Brady Kim, Bokyung Chen, Yiran Ku, Jonathan Huang, Qijia Li, Hai Zheng, Qilin
Author_xml	– sequence: 1 givenname: Bokyung orcidid: 0000-0001-5578-5237 surname: Kim fullname: Kim, Bokyung email: bk.kim@rutgers.edu organization: Department of Electrical and Computer Engineering, Rutgers, The State University of New Jersey, Piscataway, NJ, USA – sequence: 2 givenname: Qijia orcidid: 0000-0001-6073-3546 surname: Huang fullname: Huang, Qijia organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA – sequence: 3 givenname: Brady orcidid: 0000-0003-2032-0960 surname: Taylor fullname: Taylor, Brady organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA – sequence: 4 givenname: Qilin orcidid: 0009-0007-2787-9464 surname: Zheng fullname: Zheng, Qilin organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA – sequence: 5 givenname: Jonathan orcidid: 0009-0009-6690-4516 surname: Ku fullname: Ku, Jonathan organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA – sequence: 6 givenname: Yiran orcidid: 0000-0002-1486-8412 surname: Chen fullname: Chen, Yiran organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA – sequence: 7 givenname: Hai orcidid: 0000-0003-3228-6544 surname: Li fullname: Li, Hai organization: Department of Electrical and Computer Engineering, Duke University, Durham, NC, USA
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/40512641$$D View this record in MEDLINE/PubMed
BookMark	eNpdkUtLxDAQx4Movr-AiAS8eMmaV9PG21pXXVhRXMXjkrZTjXTT2qQHBb-7WXwcvMwM8_8xzx207loHCB0wOmKM6tOH83w8H3HKk5FIUs1TsYa2mZaUaK3p-ioWnMhEJltox_tXShPFNd9EW5ImjCvJttHnzdDc2TM8xjEIlpSN8R4bV-E7Eyy4QKaugg6icQFPOttAF2yJ52A_hh5wvuJtbaHHTza84LwlF-Dts4MKT103BOJDLNQ6079HcRkz1j0T68j8fnyzhzZq03jY__G76PFy8pBfk9nt1TQfz4jligcCRjEpC53qhKm0SKtaZVKKItNZaYAKYWooWEopqJTpgpa6LLioGJVQV0KB2EUn33W7vn0bwIfF0voSmsY4aAe_EJxlSkmuRUSP_6Gv7dC7OF2kpMpiA51F6uiHGoolVIuut8u44eL3sBE4_AYsAPzJ8W1CqjQRX9i6hYA
CODEN	ITBCCW
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
DBID	97E RIA RIE CGR CUY CVF ECM EIF NPM 7QO 7SP 7TB 8FD FR3 L7M P64 7X8
DOI	10.1109/TBCAS.2025.3579273
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE/IET Electronic Library (IEL) (UW System Shared) Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed Biotechnology Research Abstracts Electronics & Communications Abstracts Mechanical & Transportation Engineering Abstracts Technology Research Database Engineering Research Database Advanced Technologies Database with Aerospace Biotechnology and BioEngineering Abstracts MEDLINE - Academic
DatabaseTitle	MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) Biotechnology Research Abstracts Technology Research Database Mechanical & Transportation Engineering Abstracts Electronics & Communications Abstracts Engineering Research Database Advanced Technologies Database with Aerospace Biotechnology and BioEngineering Abstracts MEDLINE - Academic
DatabaseTitleList	Biotechnology Research Abstracts MEDLINE - Academic MEDLINE
Database_xml	– sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE/IET Electronic Library (IEL) (UW System Shared) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering
EISSN	1940-9990
EndPage	766
ExternalDocumentID	40512641 11034675
Genre	orig-research Journal Article
GrantInformation_xml	– fundername: National Science Foundation (NSF) grantid: IIS-2332744; CNS-2233808; CNS-2112562 – fundername: Rutgers University Startup Package grantid: 304291
GroupedDBID	--- 0R~ 29I 4.4 5GY 5VS 6IK 97E AAJGR AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK ACPRK AENEX AETIX AFRAH AGQYO AGSQL AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ IFIPE IPLJI JAVBF LAI M43 O9- OCL P2P RIA RIE RNS AARMG CGR CUY CVF ECM EIF NPM RIG 7QO 7SP 7TB 8FD FR3 L7M P64 7X8
ID	FETCH-LOGICAL-i262t-ea6144b9795167b7df68443b898cae033afeb1700e6719b0c9cb23d104efd36e3
IEDL.DBID	RIE
ISICitedReferencesCount	0
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001552877400010&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	1932-4545 1940-9990
IngestDate	Thu Oct 02 22:34:58 EDT 2025 Sat Nov 01 14:14:32 EDT 2025 Fri Aug 08 01:52:01 EDT 2025 Sun Sep 28 03:48:01 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	4
Language	English
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-i262t-ea6144b9795167b7df68443b898cae033afeb1700e6719b0c9cb23d104efd36e3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 content type line 23
ORCID	0009-0007-2787-9464 0000-0003-2032-0960 0000-0002-1486-8412 0000-0003-3228-6544 0000-0001-6073-3546 0000-0001-5578-5237 0009-0009-6690-4516
PMID	40512641
PQID	3246871998
PQPubID	85510
PageCount	11
ParticipantIDs	proquest_journals_3246871998 pubmed_primary_40512641 ieee_primary_11034675 proquest_miscellaneous_3218664293
PublicationCentury	2000
PublicationDate	2025-08-01
PublicationDateYYYYMMDD	2025-08-01
PublicationDate_xml	– month: 08 year: 2025 text: 2025-08-01 day: 01
PublicationDecade	2020
PublicationPlace	United States
PublicationPlace_xml	– name: United States – name: New York
PublicationTitle	IEEE transactions on biomedical circuits and systems
PublicationTitleAbbrev	TBCAS
PublicationTitleAlternate	IEEE Trans Biomed Circuits Syst
PublicationYear	2025
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
SSID	ssj0056292
Score	2.398772
Snippet	Unprovoked seizures have threatened epilepsy patients over 70 million. Automated classification to detect and predict seizures could bring seizure-free lives...
SourceID	proquest pubmed ieee
SourceType	Aggregation Database Index Database Publisher
StartPage	756
SubjectTerms	Accuracy Algorithms Artificial neural networks Biological system modeling Brain modeling Chips (memory devices) Classification Classification algorithms Co-design Computation Computer architecture computing-in-memory Convulsions & seizures cross-patient diagnosis automation EEG Electroencephalography Epilepsy Epilepsy - classification Epilepsy - diagnosis Fast Fourier transforms Hardware Humans input-stationary dataflow Multiclass seizure Neural networks Neural Networks, Computer neural processor patient-independence Quality of life Quantization (signal) seizure classification seizure detection seizure prediction Seizures Seizures - classification Seizures - diagnosis Signal Processing, Computer-Assisted SRAM-CIM Static random access memory Training
Title	MulPi: A Multi-class and Patient-Independent Epileptic Seizure Classifier With Co-Designed Input-stationary Computing-in-SRAM
URI	https://ieeexplore.ieee.org/document/11034675 https://www.ncbi.nlm.nih.gov/pubmed/40512641 https://www.proquest.com/docview/3246871998 https://www.proquest.com/docview/3218664293
Volume	19
WOSCitedRecordID	wos001552877400010&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE/IET Electronic Library (IEL) (UW System Shared) customDbUrl: eissn: 1940-9990 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0056292 issn: 1932-4545 databaseCode: RIE dateStart: 20070101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3fT9swED4BQggeNhgddGPISHs1TeMkjvfWdSAqUVRREH2rHOciIk1p1SZIIO1_39lJqr3wsLdIieLI9-u7-O47gO-E8X1MTMQJfcc88NHwRCvkjo1MmTCWmSNxvZV3d_FspiZNs7rrhUFEV3yGl_bSneWnC1PZX2U9ClWCDDvchm0po7pZq3W7FMfdBGQLSCyRd9h2yHiq9_BzOJhSLuiHlyKUiiL2PuwRUukTGug3Y1XeR5gu0lx__M9vPIQPDaRkg1oHjmALi09w8A_R4DH8GVe_J_kPNmCu4ZYbi5mZLlI2qXlV-WgzDrdkV0tyFeRKDJti_latkLnRmXlGIZQ95eUzGy74L1f6gSkbFcuq5Ov6SF-vXlk9KILW5XnBp_eDcQcer68ehje8GbzAcz_yS47apomJkgS_IpnINIviIBBJrGKj0RNCZ-TipedhJPsq8YwyiS9SyuwwS0WE4jPsFIsCT4FJVF4cpIHWQUDILVCeSCmFCYXFNonpd6Fj93C-rLk15u32deGsFce8sar1nMBfRAkeZYhduNjcJnuwhxy6wEVln7EUfhRlRRdOajFuXt6K_Ms7i36FfaskdX3fGeyUqwq_wa55KfP16pyUbhafO6X7C6TP0YQ
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3fa9swED66dmzdw36169K1mwZ7VetYtmX1LU1bGpaEsGSsb0aWz8wwnJDYhQ32v-8k22EvfeibwcYyku7uO-vu-wC-EMb3MTURJ_Qd88BHw1OtkDs2MmXCWOaOxHUsp9P47k7N2mZ11wuDiK74DM_spTvLz5amtr_KzilUCTLs8AnsWemstl2rc7wUyZ0GsoUklso77HpkPHW-uBwO5pQN-uGZCKWimL0Pzwir9AkP9FthlYcxpos1N68e-ZWv4WULKtmg2QVvYAfLt_DiP6rBA_g7qX_Nigs2YK7llhuLmpkuMzZrmFX5aCuIW7HrFTkLciaGzbH4U6-ROfHMIqcgyn4U1U82XPIrV_yBGRuVq7rim-ZQX69_s0YqgsblRcnn3waTQ_h-c70Y3vJWeoEXfuRXHLVNFFMlCYBFMpVZHsVBINJYxUajJ4TOyclLz8NI9lXqGWVSX2SU22GeiQjFO9gtlyW-ByZReXGQBVoHAWG3QHkioyQmFBbdpKbfg0M7h8mqYddIuunrwUm3HElrV5uE4F9EKR7liD34vL1NFmGPOXSJy9o-Y0n8KM6KHhw1y7h9ebfkxw8M-gme3y4m42Q8mn79APt2wzTVfiewW61rPIWn5r4qNuuPbuv9A8gi0-U
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=MulPi%3A+A+Multi-class+and+Patient-Independent+Epileptic+Seizure+Classifier+With+Co-Designed+Input-stationary+Computing-in-SRAM&rft.jtitle=IEEE+transactions+on+biomedical+circuits+and+systems&rft.au=Kim%2C+Bokyung&rft.au=Huang%2C+Qijia&rft.au=Taylor%2C+Brady&rft.au=Zheng%2C+Qilin&rft.date=2025-08-01&rft.eissn=1940-9990&rft.volume=19&rft.issue=4&rft.spage=756&rft_id=info:doi/10.1109%2FTBCAS.2025.3579273&rft_id=info%3Apmid%2F40512641&rft.externalDocID=40512641
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1932-4545&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1932-4545&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1932-4545&client=summon