OSCR

Decoding Mandarin Action Verbs from EEG Using a Dual-LSTM Network: Towards Practical Assistive Brain-Computer Interfaces.

Code ↔ Paper

The paper beside its authors' code: matches between them have not been computed for this paper yet.

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

MATLAB · 127 lines · 5.4 KB · no license

  1. %% Data format transformation
  2. clear;clc
  3. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\';
  4. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\1_set\';
  5. cd(data_path)
  6. files = dir(fullfile(data_path, '*.cnt'));
  7. fn = {files.name};
  8. for i = 1:length(fn)
  9. EEG = pop_loadcnt(char(fn(i)), 'dataformat', 'auto', 'memmapfile', '');
  10. EEG = eeg_checkset( EEG );
  11. setname = [fn{i}(1:end-4),'set'];
  12. EEG = pop_saveset( EEG, 'filename',setname,'filepath',save_path);
  13. end
  14. %% Movement artifacts were identified and removed by visual inspection
  15. clear;clc
  16. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\1_set\';
  17. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\2_artifact\';
  18. cd(data_path)
  19. files = dir('*.set');
  20. fn = {files.name};
  21. for i = 1:length(fn)
  22. EEG = pop_loadset('filename',fn{i},'filepath',data_path);
  23. EEG = eeg_checkset( EEG );
  24. EEG=pop_chanedit(EEG, 'lookup','D:\\0_toolboxes\\eeglab13_0_0b\\plugins\\dipfit2.2\\standard_BESA\\standard-10-5-cap385.elp');
  25. EEG = pop_select( EEG,'nochannel',{'HEO' 'VEO' 'CB1' 'CB2'});
  26. EEG = pop_epoch( EEG, { '1' '2' '4'}, [-1 1.5], 'newname', ' resampled epochs', 'epochinfo', 'yes');
  27. EEG = pop_saveset( EEG, 'filename',fn{i},'filepath',save_path);
  28. end
  29. % Browse each data manually pop_eegplot( EEG, 1, 1, 1);
  30. % visually inspected to remove epochs contaminated by eye movements,
  31. % electrode drifting, amplifier blocking, or EMG artifacts EEG=pop_rejepoch( EEG, [] ,0);
  33. %% Re-reference
  34. clear;clc
  35. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\2_artifact\';
  36. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\3_ref\';
  37. cd(data_path)
  38. files = dir('*.set');
  39. fn = {files.name};
  40. for i = 1:length(fn)
  41. EEG = pop_loadset('filename',fn{i},'filepath',data_path);
  42. EEG = eeg_checkset( EEG );
  43. EEG = pop_reref( EEG, [33 43] );
  44. EEG = pop_saveset( EEG, 'filename',fn{i},'filepath',save_path);
  45. end
  47. %% Filter
  48. clear;clc
  49. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\3_ref\';
  50. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\4_filter\';
  51. cd(data_path)
  52. files = dir('*.set');
  53. fn = {files.name};
  54. for i = 1:length(fn)
  55. EEG = pop_loadset('filename',fn{i},'filepath',data_path);
  56. EEG = eeg_checkset( EEG );
  57. EEG = pop_eegfiltnew(EEG, [], 0.1, 16500, true, [], 1);
  58. EEG = eeg_checkset( EEG );
  59. EEG = pop_eegfiltnew(EEG, [], 30, 220, 0, [], 1);
  60. EEG = eeg_checkset( EEG );
  61. EEG = pop_saveset( EEG, 'filename',fn{i},'filepath',save_path);
  62. end
  63. EEG = pop_interp(EEG, [33], 'spherical');
  64. EEG = eeg_checkset( EEG );
  65. EEG = pop_epoch( EEG, { '1' '2' '4' }, [-1 1.5], 'newname', 'CNT file epochs', 'epochinfo', 'yes');
  66. EEG = eeg_checkset( EEG );
  67. pop_eegplot( EEG, 1, 1, 1);
  68. EEG = eeg_checkset( EEG );
  69. EEG = pop_runica(EEG, 'extended',1,'interupt','on');
  70. EEG = eeg_checkset( EEG );
  71. pop_ADJUST_interface( );
  72. EEG = eeg_checkset( EEG );
  73. EEG = pop_subcomp( EEG, [1 2 4 32], 0);
  74. EEG = eeg_checkset( EEG );
  75. pop_eegplot( EEG, 1, 1, 1);
  76. EEG = eeg_checkset( EEG );
  77. EEG = pop_rejepoch( EEG, [32 36 38 39 41 42 45 46:50 65 66:68 76 78 79:82 84 85 90 94 95 111 115 117 119 120 128 130:2:132] ,0);
  78. EEG = pop_reref( EEG, [33 43] );
  79. EEG = eeg_checkset( EEG );
  80. EEG = pop_loadset('filename','sub1ICAout.set','filepath','C:\\Users\\X\\Desktop\\EEG results\\results\\MM\\sub1\\');
  81. EEG = eeg_checkset( EEG );
  82. EEG = pop_epoch( EEG, { '1' '2' '4' }, [-0.1 0.6], 'newname', 'CNT file epochs pruned with ICA epochs', 'epochinfo', 'yes');
  83. EEG = eeg_checkset( EEG );
  84. EEG = pop_rmbase( EEG, [-100 0]);
  85. EEG = eeg_checkset( EEG );
  86. EEG = pop_eegthresh(EEG,1,[1:60] ,-100,100,-0.1,0.598,0,0);
  87. pop_eegplot( EEG, 1, 1, 1);
  89. %% RunICA
  90. clear;clc
  91. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\4_filter\';
  92. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\5_ICA\';
  93. cd(data_path)
  94. files = dir('*.set');
  95. fn = {files.name};
  96. for i = 1:length(fn)
  97. EEG = pop_loadset('filename',fn{i},'filepath',data_path);
  98. EEG = eeg_checkset( EEG );
  99. EEG = pop_runica(EEG, 'extended',1,'interupt','on');
  100. EEG = pop_saveset( EEG, 'filename',fn{i},'filepath',save_path);
  101. end
  102. %% Identify and reject bad ingredients
  103. %Load each data manually pop_eegplot( EEG, 1, 1, 1);
  104. %the movement artifacts and eye-blink artifacts were identified and
  105. %rejected by visual inspection and Adjust pop_ADJUST_interface( );
  106. %data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\5_ICA\';
  107. %save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\6_adjust\';
  109. %% epoch extract;baseline correction;extreme value artifact removal
  110. clear;clc
  111. data_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\6_adjust\';
  112. save_path = 'E:\dataset\03_ERP_Rawdata\E04_ERP_Rawdata\preprocess\7_epoch\';
  113. cd(data_path)
  114. files = dir('*.set');
  115. fn = {files.name};
  116. for i = 1:length(fn)
  117. EEG = pop_loadset('filename',fn{i},'filepath',data_path);
  118. EEG = eeg_checkset( EEG );
  119. EEG = pop_epoch( EEG, { '1' '2' '4' }, [-0.1 0.6], 'newname', 'CNT file epochs pruned with ICA epochs', 'epochinfo', 'yes');
  120. EEG = eeg_checkset( EEG );
  121. EEG = pop_rmbase( EEG, [-100 0]);
  122. EEG = eeg_checkset( EEG );
  123. EEG = pop_eegthresh(EEG,1,[1:60] ,-100,100,-0.1,0.598,0,0);
  124. EEG = pop_saveset( EEG, 'filename',fn{i},'filepath',save_path);
  125. end
  126. %% Finalization
  127. % Load each data manually to ensure data is clean after preprocessingpop_eegplot( EEG, 1, 1, 1);

preprocessing_code_eegkab.m, no license · at the source

Overview

Authors: Binshuo Liu1, Gengbiao Chen2, Lairong Yin2, Jing Liu1
  1. International College of Engineering, Changsha University of Science and Technology, Changsha 410114, China; (B.L.); (J.L.)
  2. College of Mechanical and Vehicle Engineering, Changsha University of Science and Technology, Changsha 410114, China
Institutions: Changsha University of Science and Technology (China)
Journal: Sensors (Basel, Switzerland), volume 26, issue 9, article 2749
Dates: received 16 March 2026; accepted 26 April 2026; published online 29 April 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.3390/s26092749 · PMID 42122470 · PMCID PMC13165879 · OpenAlex W7160076964
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: EEG (modality), human (organism), cognitive (subfield)
Methods: Statistics, Machine learning
Keywords: EEG signal processing, Mandarin verb decoding, recurrent neural networks (RNN), brain–computer interfaces (BCI), assistive technology
MeSH: Brain-Computer Interfaces*, Electroencephalography*, Language*, Adult, Humans, Long Short Term Memory, Male, Neural Networks, Computer, Signal Processing, Computer-Assisted, Speech, Support Vector Machine (* major topic)
Topic: EEG and Brain-Computer Interfaces (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: Science and Technology Innovation Program of Hunan Province (2024RC1054, 2025JJ70113); Hunan Provincial Natural Science Foundation of China (2025JJ70123); National Natural Science Foundation of China
Citations: not cited yet (Europe PMC); 48 references in the paper

Abstract

Electroencephalogram (EEG)-based brain–computer interfaces (BCIs) offer a promising pathway for restoring communication. Decoding tonal languages like Mandarin from EEG remains challenging due to homophones and complex temporal dynamics. This study investigates the decoding of six high-frequency Mandarin action verbs—Chi (eat), He (drink), Chuan (wear), Na (take), Kan (look), and Dai (put on)—from EEG signals. We designed a visual-cue-based overt speech production experiment and collected EEG data from 30 participants during visually guided verb reading aloud. A recurrent neural network framework incorporating dual Long Short-Term Memory (LSTM) layers was implemented to model the long-range temporal dependencies in EEG patterns. The proposed model was compared against a traditional Common Spatial Pattern combined with Support Vector Machine (CSP-SVM) baseline. Our LSTM-based model achieved an average classification accuracy of 69.93% ± 3.07% for the six-class task, significantly outperforming the CSP-SVM baseline (36.53% ± 3.17%). Accuracy exceeded 75% under specific training conditions, including more than 15 training repetitions and a training-data proportion of 38%. Furthermore, the model attained this performance level utilizing approximately 38% of the available trial data for training, demonstrating data efficiency. The results indicate that the LSTM architecture can effectively capture the neural signatures associated with Mandarin verb processing, providing a foundation for developing practical EEG-based assistive communication technologies. The inference latency of the trained model, quantified as the post-training per-trial testing time, was under 2 s, supporting near-real-time applications.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above.

OSF nmke5

License: none: the authors keep all their rights
State: the link answers, verified on 30 September 2026
Evidence: files inventoried
Languages: MATLAB (2), R (1)
Size: 23 files, 3 scripts
Software Heritage: not checked
Found in: the text, “5.1. Experimental Subjects”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: EEGLAB (2 files), FieldTrip (1 file), ggplot2 (1 file), ggpubr (1 file), lme4 (1 file), lmerTest (1 file), Statistics and Machine Learning Toolbox (1 file), tidyverse (1 file)
Availability: 1 check, the latest on 30 September 2026: the link answers (HTTP 200)
  • 30 September 2026: the link answers (HTTP 200)
3 files
At the source: doi.org/10.17605/OSF.IO/NMKE5

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 3 scripts, each with its path and the digest of its content;
  • no match between paragraphs and code yet;
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data Availability Statement

To ensure full reproducibility, the complete preprocessed EEG dataset and all analysis code have been deposited in a public repository [38]. The shared data comprises the EEG signals after standard preprocessing steps, which serves as the direct input to the decoding models. This level of data sharing aligns with best practices for enabling independent verification and extension of the analytical findings presented in this work.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 30 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 4 authors, 5 keywords, 11 MeSH terms, 3 funders, 48 references.

Cite

This paper

Liu, B., Chen, G., Yin, L., & Liu, J. (2026). Decoding Mandarin Action Verbs from EEG Using a Dual-LSTM Network: Towards Practical Assistive Brain-Computer Interfaces. Sensors (Basel, Switzerland), 26(9), 2749. https://doi.org/10.3390/s26092749

BibTeX

@article{liu2026decoding,
author = {Liu, Binshuo and Chen, Gengbiao and Yin, Lairong and Liu, Jing},
title = {{Decoding Mandarin Action Verbs from EEG Using a Dual-LSTM Network: Towards Practical Assistive Brain-Computer Interfaces}},
journal = {Sensors (Basel, Switzerland)},
year = {2026},
month = apr,
volume = {26},
number = {9},
pages = {2749},
publisher = {Multidisciplinary Digital Publishing Institute (MDPI)},
issn = {1424-8220},
doi = {10.3390/s26092749},
url = {https://doi.org/10.3390/s26092749},
pmid = {42122470},
pmcid = {PMC13165879}
}

RIS

TY - JOUR
AU - Liu, Binshuo
AU - Chen, Gengbiao
AU - Yin, Lairong
AU - Liu, Jing
TI - Decoding Mandarin Action Verbs from EEG Using a Dual-LSTM Network: Towards Practical Assistive Brain-Computer Interfaces
T2 - Sensors (Basel, Switzerland)
J2 - Sensors (Basel)
PY - 2026
DA - 2026/04/29
VL - 26
IS - 9
SP - 2749
SN - 1424-8220
PB - Multidisciplinary Digital Publishing Institute (MDPI)
DO - 10.3390/s26092749
UR - https://doi.org/10.3390/s26092749
LA - en
ER -

CSL-JSON

{
"id": "10.3390/s26092749",
"type": "article-journal",
"title": "Decoding Mandarin Action Verbs from EEG Using a Dual-LSTM Network: Towards Practical Assistive Brain-Computer Interfaces",
"container-title": "Sensors (Basel, Switzerland)",
"author": [
{
"family": "Liu",
"given": "Binshuo"
},
{
"family": "Chen",
"given": "Gengbiao"
},
{
"family": "Yin",
"given": "Lairong"
},
{
"family": "Liu",
"given": "Jing"
}
],
"container-title-short": "Sensors (Basel)",
"volume": "26",
"issue": "9",
"page": "2749",
"DOI": "10.3390/s26092749",
"PMID": "42122470",
"PMCID": "PMC13165879",
"ISSN": "1424-8220",
"publisher": "Multidisciplinary Digital Publishing Institute (MDPI)",
"URL": "https://doi.org/10.3390/s26092749",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
29
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1523/eneuro.0316-25.2026 [code]
Neural Mechanisms of Self-Generated Action Sequences.
Journal: eNeuro
In common: EEGLAB, FieldTrip, lmerTest, 4 other tools, EEG, cognitive
[2] doi:10.1093/cercor/bhag077 [code]
The longitudinal development of intrinsic timescales in infancy and their relation to alpha brain rhythm.
Journal: Cerebral cortex (New York, N.Y. : 1991)
In common: EEGLAB, FieldTrip, lmerTest, 4 other tools, EEG
[3] doi:10.1016/j.neuroimage.2026.122115 [code]
Midfrontal theta power relates to response speeding following frustrative nonreward.
Journal: NeuroImage
In common: EEGLAB, lmerTest, lme4, 4 other tools, EEG, cognitive
[4] doi:10.1111/psyp.70265 [code]
Neurocognitive Dynamics of Translating Information From a Spatial Map Into Action.
Journal: Psychophysiology
In common: EEGLAB, lmerTest, lme4, 4 other tools, EEG, cognitive
[5] doi:10.1371/journal.pbio.3003979 [code]
Impaired midfrontal‑motor theta phase synchronization characterizes maladaptive motivational behavior in people with obsessive‑compulsive disorder.
Journal: PLoS biology
In common: EEGLAB, FieldTrip, lmerTest, 4 other tools, EEG
[6] doi:10.7554/elife.107088 [code]
Development of auditory and spontaneous movement responses to music over the first postnatal year.
Journal: eLife
In common: EEGLAB, FieldTrip, lme4, 4 other tools, EEG
[7] doi:10.1093/braincomms/fcag130 [code]
Sleep increases firing rate modulation during interictal epileptiform discharges in mesial temporal structures.
Journal: Brain communications
In common: FieldTrip, lmerTest, lme4, 4 other tools, EEG
[8] doi:10.1038/s41467-026-75799-8 [code]
Behaviourally driven closed-loop beta-tACS enhances beta activity and motor behaviour.
Journal: Nature communications
In common: EEGLAB, FieldTrip, lmerTest, 4 other tools
[9] doi:10.1162/imag.a.1245 [code]
Towards precision EEG connectomics: Evaluating the benefits of dense sampling.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: FieldTrip, lmerTest, lme4, 4 other tools, EEG
[10] doi:10.64898/2026.05.08.26348885 [code]
Insights from nine nights of self-applied, low-density sleep EEG during sleep restriction therapy: a proof-of-concept evaluation
Journal: medRxiv (preprint)
In common: FieldTrip, lmerTest, lme4, 4 other tools, EEG

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.