IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2022-03-02 13:25
[Poster Presentation] Transformer-based Text Decoding using Electrocorticography
Shuji Komeiji, Kai Shigemi (TUAT), Takumi Mitsuhashi, Yasushi Iimura, Hiroharu Suzuki, Hidenori Sugano (Juntendo Univ.), Koichi Shinoda (Tokyo Tech), Toshihisa Tanaka (TUAT) EA2021-87 SIP2021-114 SP2021-72
Abstract (in Japanese) (See Japanese page) 
(in English) Invasive brain-machine interfaces (BMIs) are a promising neurotechnology for achieving direct speech communication from a human brain but face many challenges. This paper measured the invasive electrocorticogram (ECoG) signals from seven participating epilepsy patients as they spoke a sentence consisting of multiple phrases. A Transformer encoder was incorporated into a “sequence-to-sequence” model (Transformer Seq2seq) to decode spoken sentences from the ECoG. A Transformer is a successful neural network model for natural language processing and automatic speech recognition. The decoding test revealed that the use of the Transformer model achieved a minimum phrase error rate of 16.4% for one best participant; moreover, the median (±standard deviation) of PER for the Transformer Seq2seq across seven participants was31.3% (±10.0%). This result showed that the Transformer Seq2seq effectively decoded from ECoG.
Keyword (in Japanese) (See Japanese page) 
(in English) Electrocorticography / Brain--machine interface / Transformer encoder / Sequence to sequence / / / /  
Reference Info. IEICE Tech. Rep., vol. 121, no. 384, SIP2021-114, pp. 146-151, March 2022.
Paper # SIP2021-114 
Date of Issue 2022-02-22 (EA, SIP, SP) 
ISSN Online edition: ISSN 2432-6380
Copyright
and
reproduction
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF EA2021-87 SIP2021-114 SP2021-72

Conference Information
Committee EA SIP SP IPSJ-SLP  
Conference Date 2022-03-01 - 2022-03-02 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To SIP 
Conference Code 2022-03-EA-SIP-SP-SLP 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Transformer-based Text Decoding using Electrocorticography 
Sub Title (in English)  
Keyword(1) Electrocorticography  
Keyword(2) Brain--machine interface  
Keyword(3) Transformer encoder  
Keyword(4) Sequence to sequence  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Shuji Komeiji  
1st Author's Affiliation Tokyo University of Agriculture and Technology (TUAT)
2nd Author's Name Kai Shigemi  
2nd Author's Affiliation Tokyo University of Agriculture and Technology (TUAT)
3rd Author's Name Takumi Mitsuhashi  
3rd Author's Affiliation Juntendo University (Juntendo Univ.)
4th Author's Name Yasushi Iimura  
4th Author's Affiliation Juntendo University (Juntendo Univ.)
5th Author's Name Hiroharu Suzuki  
5th Author's Affiliation Juntendo University (Juntendo Univ.)
6th Author's Name Hidenori Sugano  
6th Author's Affiliation Juntendo University (Juntendo Univ.)
7th Author's Name Koichi Shinoda  
7th Author's Affiliation Tokyo Institute of Technology (Tokyo Tech)
8th Author's Name Toshihisa Tanaka  
8th Author's Affiliation Tokyo University of Agriculture and Technology (TUAT)
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
Speaker Author-1 
Date Time 2022-03-02 13:25:00 
Presentation Time 120 minutes 
Registration for SIP 
Paper # EA2021-87, SIP2021-114, SP2021-72 
Volume (vol) vol.121 
Number (no) no.383(EA), no.384(SIP), no.385(SP) 
Page pp.146-151 
#Pages
Date of Issue 2022-02-22 (EA, SIP, SP) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan