Analytical algorithm of SSVEP based on improved convolutional neural network

doi:10.3969/j.issn.1000-1565.2025.05.009

Abstract

Abstract: Steady-state visual evoked potentials(SSVEP)is a commonly used modality in brain-computer interfaces, which usually suffers from insufficient utilization of time domain and frequency information, resulting in imprecise and untimely signal resolution. For this reason, this paper proposes an improved SSVEP analysis algorithm for convolutional neural network model. A multi-channel input model is designed. With multiple frequency band filtered signals as input, the depth features of time domain and frequency domain are extracted by parallel time attention module and multi-band combination module.The classification module realizes the accurate analysis of SSVEP signal through the fusion analysis of multi-feature domain. The proposed algorithm is verified on two common data sets, and the classification accuracy- DOI:10.3969/j.issn.1000-1565.2025.05.009基于改进卷积神经网络的SSVEP解析算法杨建利^1,2,赵松磊¹,刘凤双¹,杨晓茹³,张烁⁴(1.河北大学电子信息工程学院,河北保定 071002;2.河北省数字医疗工程重点实验室,河北保定 071002;3.保定市生产力促进中心,河北保定 071000;4.河北大学附属医院人事处,河北保定 071000)摘要:稳态视觉诱发电位(steady-state visual evoked potentials,SSVEP)是脑机接口中常用的一种方式,该方式通常存在时域和频域信息利用不充分,导致信号解析不精准、不及时的问题.为此,本文提出了一种改进卷积神经网络模型的SSVEP解析算法.设计了多通道输入模型,以多个频带滤波的信号作为输入,利用并行的时间注意力模块和多频带组合模块分别提取时域和频域的深度特征,经分类模块的多特征域融合分析实现了SSVEP信号的精准解析.在2个公共数据集上对本文算法进行了验证,分别取得了98.14%和82.72%的分类准确率.实验结果表明,该模型具有较高性能和鲁棒性,有助于推动基于 SSVEP 的脑机接口发展.关键词:卷积神经网络;脑机接口;稳态视觉诱发电位中图分类号:TP391 文献标志码:A 文章编号:1000-1565(2025)05-0530-11Analytical algorithm of SSVEP based on improved convolutional neural networkYANG Jianli^1,2, ZHAO Songlei¹, LIU Fengshuang¹, YANG Xiaoru³, ZHANG Shuo⁴(1. College of Electronic and Information Engineering, Hebei University, Baoding 071002, China; 2. Key Laboratory of Digital Medical Engineering of Hebei Province, Baoding 071002, China; 3. Productivity Promotion Center of Baoding,Baoding 071000, China; 4. Human Resources Department, Affiliated Hospital of Hebei University, Baoding 071000, China)Abstract: Steady-state visual evoked potentials(SSVEP)is a commonly used modality in brain-computer interfaces, which usually suffers from insufficient utilization of time domain and frequency information, resulting in imprecise and untimely signal resolution. For this reason, this paper proposes an improved SSVEP analysis algorithm for convolutional neural network model. A multi-channel input model is designed. With multiple frequency band filtered signals as input, the depth features of time domain and frequency domain are extracted by parallel time attention module and multi-band combination module.The classification module realizes the accurate analysis of SSVEP signal through the fusion analysis of multi-feature domain. The proposed algorithm is verified on two common data sets, and the classification accuracy- 收稿日期:2024-09-26;修回日期:2025-04-22 基金项目:河北大学自然科学多学科交叉研究计划项目(DXK202205) 第一作者:杨建利(1987—),男,河北大学副教授,博士,主要从事脑科学领域的信号处理与建模研究.E-mail:yangjianli_1987@126.com 通信作者:张烁(1988—),女,河北大学附属医院经济师,主要从事数据统计、智能分析方向研究.E-mail:Zhangshuolele@126.com第5期杨建利等:基于改进卷积神经网络的SSVEP解析算法河北大学学报(自然科学版) 第45卷is 98.14% and 82.72%, respectively. The experimental results show that the model exhibits high performance and robustness, thereby facilitating the development of brain-computer interface based on SSVEP.

Key words: convolutional neural network, brain-computer interface, SSVEP

CLC Number:

TP391

YANG Jianli, ZHAO Songlei, LIU Fengshuang, YANG Xiaoru, ZHANG Shuo. Analytical algorithm of SSVEP based on improved convolutional neural network[J]. Journal of Hebei University(Natural Science Edition), 2025, 45(5): 530-540.

References

[1] WOLPAW J R, BIRBAUMER N, MCFARLAND D J, et al. Brain-computer interfaces for communication and control[J]. Clin Neurophysiol, 2002, 113(6): 767-791. DOI: 10.1016/s1388-2457(02)00057-3.
[2] CHEN X, ZHAO B, WANG Y, et al. Control of a 7-DOF robotic arm system with an SSVEP-based BCI[J]. IntJ Neural Syst, 2018, 28(8): 1850018. DOI: 10.1142/s0129065718500181.
[3] LI Y, PAN J, WANG F, et al. A hybrid BCI system combining P300 and SSVEP and its application to wheelchair control[J]. IEEE T Biomed Eng, 2013, 60(11): 3156-3166. DOI: 10.1109/tbme.2013.2270283.
[4] WONG C M, TANG Q, DA CRUZ J N, et al. A multi-channel SSVEP-based BCI for computer games with analogue control[C] //2015 CIVEMSA. IEEE, 2015: 1-6. DOI: 10.1109/CIVEMSA.2015.7158612.
[5] COOGAN C G, HE B. Brain-computer interface control in a virtual reality environment and applications for the internet of things[J]. IEEE Access, 2018, 6: 10840-10849. DOI: 10.1109/access.2018.2809453.
[6] MAO X, LI W, LEI C, et al. A brain-robot interaction system by fusing human and machine intelligence[J]. IEEE T Neur Sys Reh, 2019, 27(3): 533-542. DOI: 10.1109/tnsre.2019.2897323.
[7] ZHANG Y, XU P, LIU T, et al. Multiple frequencies sequential coding for SSVEP-based brain-computer interface[J]. PLoS One, 2012, 7(3): e29519. DOI: 10.1371/journal.pone.0029519.
[8] ZHAO D, WANG T, TIAN Y, et al. Filter Bank Convolutional Neural Network for SSVEP Classification[J]. IEEE Access, 2021, 9: 147129-147141. DOI: 10.1109/access.2021.3124238.
[9] 林中林,张雅静,高小榕, 等.一种通过脑电信号实时检测双眼竞争的方法[J].航天医学与医学工程,2007(5):381-384.DOI:10.16289/j.cnki.1002-0837.2007.05.007. DOI: 10.16289/j.cnki.1002-0837.2007.05.007.
[10] ZHANG Y, YIN E, LI F, et al. Hierarchical feature fusion framework for frequency recognition in SSVEP-based BCIs[J]. Neural Networks, 2019, 119: 1-9. DOI: 10.1016/j.neunet.2019.07.007.
[11] SHAO X, LIN M. Filter bank temporally local canonical correlation analysis for short time window SSVEPs classification[J]. Cogn Neurodynamics, 2020, 14(5): 689-696. DOI: 10.1007/s11571-020-09620-7.
[12] WONG C M, WAN F, WANG B, et al. Learning across multi-stimulus enhances target recognition methods in SSVEP-based BCIs[J]. J Neural Eng, 2020, 17(1): 016026. DOI: 10.1088/1741-2552/ab2373.
[13] 杜光景,谢俊,张玉彬, 等. 用于稳态视觉诱发电位脑机接口目标识别的深度学习方法[J]. 西安交通大学学报, 2019,53(11):42-48. DOI: 10.7652/xjtuxb201911006.
[14] RAVI A, BENI N H, MANUEL J, et al. Comparing user-dependent and user-independent training of CNN for SSVEP BCI[J]. J Neural Eng, 2020, 17(2): 026028. DOI: 10.1088/1741-2552/ab6a67.
[15] DANG W, LI M, LV D, et al. MHLCNN: Multi-harmonic linkage CNN model for SSVEP and SSMVEP signal classification[J]. IEEE T Circuits-II, 2021, 69(1): 244-248. DOI: 10.1109/tcsii.2021.3091803.
[16] GUNEY O B, OBLOKULOV M, OZKAN H. A deep neural network for ssvep-based brain-computer interfaces[J]. IEEE T Biomed Eng, 2021, 69(2): 932-944. DOI: 10.1109/tbme.2021.3110440.
[17] GUNEY O B, OBLOKULOV M, OZKAN H. A deep neural network for ssvep-based brain-computer interfaces[J]. IEEE T Biomed Eng, 2021, 69(2): 932-944. DOI: 10.1109/TBME.2021.3110440.
[18] WANG Y, CHEN X, GAO X, et al. A benchmark dataset for SSVEP-based brain-computer interfaces[J]. IEEE T Neur Sys Reh, 2016, 25(10): 1746-1752. DOI: 10.1109/tnsre.2016.2627556.
[19] LIU B, HUANG X, WANG Y, et al. BETA: A large benchmark database toward SSVEP-BCI application[J]. Front Neurosci, 2020, 14: 627. DOI: 10.3389/fnins.2020.00627.
[20] MÜLLER-PUTZ G R, SCHERER R, Brauneis C, et al. Steady-state visual evoked potential(SSVEP)-based communication: Impact of harmonic frequency components[J]. J Neural Eng, 2005, 2(4): 123. DOI: 10.1088/1741-2560/2/4/008.
[21] WANG Y, GAO X, HONG B, et al. Brain-computer interfaces based on visual evoked potentials[J]. IEEE Eng Med Biol, 2008, 27(5): 64-71. DOI: 10.1109/memb.2008.923958.
[22] YIN E, ZHOU Z, JIANG J, et al. A dynamically optimized SSVEP brain-computer interface(BCI)speller[J]. IEEE T Biomed Eng, 2014, 62(6): 1447-1456. DOI: 10.1109/tbme.2014.2320948.
[23] CHEN X, WANG Y, NAKANISHI M, et al. High-speed spelling with a noninvasive brain-computer interface[J]. P Natl Acad Sci, 2015, 112(44): E6058-E6067. DOI: 10.1073/pnas.1508080112.
[24] TANAKA H, KATURA T, SATO H. Task-related component analysis for functional neuroimaging and application to near-infrared spectroscopy data[J]. NeuroImage, 2013, 64: 308-327. DOI: 10.1016/j.neuroimage.2012.08.044.
[25] ZHANG Y, YIN E, LI F, et al. Two-stage frequency recognition method based on correlated component analysis for SSVEP-based BCI[J]. IEEE T Neur Sys Reh, 2018, 26(7): 1314-1323. DOI: 10.1109/tnsre.2018.2848222.
[26] CHEN X, WANG Y, GAO S, et al. Filter bank canonical correlation analysis for implementing a high-speed SSVEP-based brain-computer interface[J]. J Neural Eng, 2015, 12(4): 046008. DOI: 10.1088/1741-2560/12/4/046008.
[27] WANG Y, NAKANISHI M, WANG Y T, et al. Enhancing detection of steady-state visual evoked potentials using individual training data[C] //2014 EMBC, IEEE, 2014: 3037-3040. DOI: 10.1109/embc.2014.6944263. (