在复杂且安全性要求较高的人机系统（如飞机驾驶）中，准确监测和识别操作员的认知负荷对于保证人机交互任务正确执行、维护操作员的健康和安全、防止事故发生至关重要。人机交互任务中的认知负荷，是在执行特定任务过程中大脑资源的占用程度或者心理压力的主观感受程度。认知负荷过高或过低都会显著影响特定任务中的人员绩效，引起人类认知行为发生变化。与传统的主观量表、任务绩效等方法相比，基于脑电（Electroencephalography，EEG）的认知负荷识别更加客观、稳定、可靠，现已得到广泛的关注。其是指通过统计学习和机器学习方法检测能表征认知负荷水平高低的EEG指标。由于EEG信号的时域非平稳、任务敏感性、被试差异性等特性，目前通用的跨任务（即适用于不同认知任务）\跨被试（即适用于不同被试）的认知负荷识别模型还存在泛化表现差、准确率低的问题，限制了预先训练模型扩展到实际场景下新认知状态\新被试的推广性。实现跨任务\跨被试认知负荷评估是从实验室走向实际应用必要的步骤，因此对认知负荷模型的泛化性开展研究具有重要意义。

本文致力于结合EEG分析和机器学习子领域方法（主要是领域自适应）构建面向认知负荷识别的泛化模型，来解决跨任务和跨被试的模型泛化性差的问题。本文围绕任务间表征不稳定、被试个体差异大、类别结构信息未充分利用、目标域数据不可用等问题展开研究。针对以上问题分别构建相应的跨任务认知负荷识别模型和跨被试认知负荷识别模型。本文的主要研究工作及主要贡献为：

针对任务间表征不稳定的问题，本文提出了基于典型迁移学习方法构建的跨任务认知负荷识别模型，解决了跨任务识别性能差的问题，实现了跨任务识别性能的提升。首先设计了一个细粒度的实验诱发范式来采集工作记忆任务和数学计算任务的EEG数据。然后提取5个频段的功率谱密度和谱相干特征作为EEG特征。进而首次尝试使用迁移成分分析、联合迁移适配等典型迁移学习模型优化学习特征并对齐特征分布，构建跨任务迁移框架来减少两个认知任务之间的特征分布差异。最后，基于支持向量机分类器对低和高认知负荷水平进行分类，不同设置下跨任务平均识别准确性提升3%至8%。

针对被试个体差异大的问题，本文提出了联合浅层特征分布对齐和深层特征域适应的跨被试认知负荷识别模型，解决了跨被试识别泛化性差的问题，实现了跨被试识别性能的提升。模型以端到端的方式直接使用原始EEG时间序列作为模型输入，并基于特征分布对齐、对抗域适应和全连接神经网络构建跨被试模型。模型通过添加分布差异度量来对齐浅层特征分布偏移，结合深层特征对抗训练来学习域不变特征，减少域间差异。与现有的无监督领域自适应方法相比，结合浅层和深层对抗性训练来学习不变的EEG特征。在自采记忆任务和公开数据集的低和高认知负荷识别实验中，该模型跨被试认知负荷识别结果提升了5%至9%。

针对类别结构信息未充分利用的问题，本文提出了联合类别信息和域信息对齐的跨被试认知负荷识别模型，解决了跨被试识别的标签偏移问题，实现了跨被试识别性能的提升。首先通过引入双分类器学习考虑类别间的相似信息，通过衡量两个分类器的预测差异考虑类别间的差异信息。其次通过域判别对抗学习考虑全局域的信息。联合局部类别信息和全局域的信息，可以更好地构建跨被试认知负荷识别模型。在自采记忆任务和公开数据集的低和高认知负荷识别实验中，该模型的跨被试认知负荷识别结果提升了2%到12%。

针对新被试数据不能参与模型训练的情况，本文提出了结合卷积神经网络和域泛化方法的跨被试认知负荷识别模型，解决了跨被试识别的测试数据不可知问题，实现了跨被试识别性能的提升。相比领域自适应技术，领域泛化的优势在于其不需要目标域数据参与模型训练，因此更适用于真实情景。模型考虑到EEG数据的时间和空间表征，同时从领域泛化的角度减少源域数据和目标域数据的分布差异，可以有效保证获取EEG数据表征的泛化性，从而提高新操作员的认知负荷识别准确性。在公开数据集的低、中、高认知负荷三分类识别实验中，所提模型取得了稳健的跨被试认知负荷识别结果，平均识别准确性提高了3%至8%。

In safety-critical human-machine systems, such as aircraft operation, the accurate monitoring and identifying of operator cognitive workload has become increasingly crucial. This is essential to ensure the proper execution of human-machine interaction tasks, as well as to maintain operator health and safety and prevent accidents. Cognitive workload in human-computer interaction tasks, which is defined as the consumption or occupation of cognitive resources or the level of psychological pressure subjectively felt during the execution of specific tasks. It is noteworthy that excessive or insufficient cognitive workload can significantly impact personnel performance and result in changes in human cognitive behavior. Among traditional evaluation methods such as subjective scales and task performance, electroencephalography (EEG)-based cognitive workload recognition is considered more objective, stable, and reliable, and has gained widespread attention. It pertains to the identification of EEG indicators that can represent cognitive workload levels through the use of statistical and machine learning techniques. However, due to the non-stationary nature of EEG signals, task sensitivity, and subject variability, the current universal cross-task (meaning models apply to different cognitive tasks) and/or cross-subject (meaning models apply to different subjects) cognitive workload recognition models face challenges with poor generalization performance and low accuracy. This limitation hinders the generalization of pre-trained models to new cognitive states or new subjects in actual scenarios. Therefore, implementing cross-task/cross-subject cognitive workload recognition is a necessary step from the laboratory toward practical application. Researching the generalization of cognitive workload models is of great significance.

This paper is committed to developing generalized models for cognitive workload recognition through the combination of EEG analysis and the subfield of machine learning. The aim is to address the issue of poor generalization of cross-task and cross-subject models. This article focuses on issues such as unstable representation between tasks, large differences between subjects, insufficient utilization of category structure information, and unavailability of target domain data. We further construct the corresponding cross-task and cross-subject cognitive workload recognition models. The main research work and contributions of this article are as follows:

Aiming at the problem of unstable representation between tasks, this paper proposes a cross-task cognitive workload recognition model based on typical transfer learning methods, which solves the problem of poor cross-task recognition performance and improves cross-task recognition performance.  Firstly, a fine-grained multi-cognitive workload task elicitation paradigm is designed to collect EEG data from two types of cognitive tasks, including the working memory task and the mathematical computing task. Then the power spectral density and coherence features of five frequency bands are extracted as EEG features. Then, for the first time, we try to use typical transfer learning models such as transfer component analysis and joint transfer matching to optimize learning features and align feature distribution, and build a cross-task transfer framework to reduce the difference in feature distribution between two cognitive tasks. Finally, the support vector machine classifier is used to classify low and high cognitive workload levels, and the average recognition accuracy of cross-task cognitive workload recognition has been improved by 3% to 8% under different settings.

Aiming at the problem of significant differences across subjects, this paper proposes a cross-subject cognitive workload recognition model that employs joint shallow feature distribution alignment and deep feature domain adaptation. The model is used to solve the problem of poor generalization performance and improves cross-subject recognition performance. The model directly uses the original EEG time series as model input in an end-to-end manner and constructs a cross-subject model based on feature distribution alignment, adversarial domain adaptation, and fully-connected neural networks. To align shallow feature distribution offsets, the model adds distribution difference measures and combines deep feature adversarial training to learn domain invariant features and reduce EEG distribution differences. Compared with existing unsupervised domain adaptation methods, the proposed approach combines shallow and deep adversarial training to learn the invariant EEG features. Cognitive workload recognition experiments are conducted on one self-collected memory task and one public dataset, and the proposed approach achieved robust cross-subject cognitive workload recognition results with the average recognition accuracy improved by 3% to 8%.

This paper proposes a novel cross-subject cognitive workload recognition model that jointly aligns category-wise and domain-wise information to address the problem of insufficient utilization of category information and label shift, resulting in improved cross-subject recognition performance. The proposed method employs bi-classifiers to learn and consider similarity information between categories while measuring prediction differences between the two classifiers to consider difference information between categories. Furthermore, adversarial learning via domain discriminator considers the information of the entire global area. By combining local category information and global domain information, the proposed method constructs a better cross-subject cognitive workload recognition model. The method achieves relatively ideal cross-subject cognitive workload recognition results for recognizing low and high cognitive workloads on one self-collected memory task and one public dataset, with the mean accuracy improving by 2% to 12%.

To address the problem of new subject data that cannot participate in model training, in this study, we propose a novel cross-subject cognitive workload recognition model that combines Convolutional Neural Network (CNN) and domain generalization. The model effectively solves the generalization problem of unseen target data and improves cross-subject recognition performance. Compared to domain adaptation technology, domain generalization has the advantage of not requiring target domain data to participate in model training, making it more suitable for real-world scenarios. To enhance the generalization of EEG data representation, we consider both temporal and spatial representation. Furthermore, we reduce the distribution differences between source domain data and target domain data from the perspective of domain generalization. This approach significantly improves the accuracy of cognitive workload recognition for new operators, achieving acceptable cross-subject cognitive workload recognition results for recognizing low, medium, and high cognitive workloads on one public EEG dataset, with the mean accuracy improving by 3% to 8%.

[1] Oluwafemi FA, Abdelbaki R, Lai JC, et al. A review of astronaut mental health in manned missions: potential interventions for cognitive and mental health challenges[J]. Life Sciences in Space Research, 2021, 28: 26-31.

[2] Ji H, Chen R, Lu L, et al. Pilot workload investigation for rotorcraft operation in low-altitude atmospheric turbulence[J]. Aerospace Science and Technology, 2021, 111(1): 106567.

[3] Cui Y, Xu Y, Wu D. EEG-based driver drowsiness estimation using feature weighted episodic training[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(11): 2263-2273.

[4] Zhang J, Liu S, Feng Q, et al. Correlative evaluation of mental and physical workload of laparoscopic surgeons based on surface electromyography and eye-tracking signals[J]. Scientific reports, 2017, 7(1): 11095.

[5] Lin Y, Deng L, Chen Z, et al. A real-time ATC safety monitoring framework using a deep learning approach[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 21: 4572-4581.

[6] Kakkos I, Dimitrakopoulos G N, Gao L, et al. Mental workload drives different reorganizations of functional cortical connectivity between 2D and 3D simulated flight experiments[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(9): 1704-1713.

[7] Babiloni F. Mental workload monitoring: new perspectives from neuroscience[C]//Human Mental Workload: Models and Applications: Third International Symposium, H-WORKLOAD 2019, Rome, Italy, November 14–15, 2019, Proceedings 3. Springer International Publishing, 2019: 3-19.

[8] 许子明, 牛一帆, 温旭云, 等. 基于脑电信号的认知负荷评估综述[J]. 航天医学与医学工程, 2021, 34(4): 339-348.

[9] Gasper K, Hackenbracht J. Too busy to feel neutral: reducing cognitive resources attenuates neutral affective states[J]. Motivation and Emotion, 2015, 39: 458-466.

[10] Aricò P, Borghini G, Di Flumeri G, et al. Passive BCI beyond the lab: current trends and future directions[J]. Physiological measurement, 2018, 39(8): 08TR02.

[11] Heard J, Harriott C E, Adams J A. A survey of workload assessment algorithms[J]. IEEE Transactions on Human-Machine Systems, 2018, 48(5): 434-451.

[12] Tao D, Tan H, Wang H, et al. A systematic review of physiological measures of mental workload[J]. International Journal of Environmental Research and Public Health, 2019, 16(15): 2716.

[13] Charles R L, Nixon J. Measuring mental workload using physiological measures: A systematic review[J]. Applied Ergonomics, 2019, 74: 221-232.

[14] Debie E, Rojas R F, Fidock J, et al. Multimodal fusion for objective assessment of cognitive workload: A review[J]. IEEE Transactions on Cybernetics, 2021, 51(3): 1542-1555.

[15] Borghini G, Astolfi L, Vecchiato G, et al. Measuring neurophysiological signals in aircraft pilots and car drivers for the assessment of mental workload, fatigue and drowsiness[J]. Neuroscience and Biobehavioral Reviews, 2014, 44: 58-75.

[16] Langer N, Von Bastian C C, Wirz H, et al. The effects of working memory training on functional brain network efficiency[J]. Cortex, 2013, 49(9): 2424-2438.

[17] Cai Y, She Q, Ji J, et al. Motor imagery EEG decoding using manifold embedded transfer learning[J]. Journal of Neuroscience Methods, 2022, 370: 109489.

[18] Hsu S H, Lin Y, Onton J, et al. Unsupervised learning of brain state dynamics during emotion imagination using high-density EEG[J]. Neuroimage, 2022, 249: 118873.

[19] Zhang J, Yin Z, Wang R. Recognition of mental workload levels under complex human–machine collaboration by using physiological features and adaptive support vector machines[J]. IEEE Transactions on Human-Machine Systems, 2014, 45(2): 200-214.

[20] Tao J, Yin Z, Liu L, et al. Individual-specific classification of mental workload levels via an ensemble heterogeneous extreme learning machine for EEG modeling[J]. Symmetry, 2019, 11(7): 944.

[21] Zhang P, Wang X, Chen J, et al. Feature weight driven interactive mutual information modeling for heterogeneous bio-signal fusion to estimate mental workload[J]. Sensors, 2017, 17(10): 2315.

[22] Kwak Y, Kong K, Song W J, et al. Multilevel feature fusion with 3d convolutional neural network for EEG-based workload estimation[J]. IEEE Access, 2020, 8: 16009-16021.

[23] Ke Y, Qi H, He F, et al. An EEG-based mental workload estimator trained on working memory task can work well under simulated multi-attribute task[J]. Frontiers in Human Neuroscience, 2014, 8: 703.

[24] Appel T, Sevcenko N, Wortha F, et al. Predicting cognitive load in an emergency simulation based on behavioral and physiological measures[C]//2019 International Conference on Multimodal Interaction (ACMMM 2019). 2019: 154-163.

[25] Appel T, Gerjets P, Hoffman S, et al. Cross-task and cross-participant classification of cognitive load in an emergency simulation game[J]. IEEE Transactions on Affective Computing, 2021, doi:10.1109/TAFFC.2021.3098237.

[26] Subha D P, Joseph P K, Acharya U R, et al. EEG signal analysis: a survey[J]. Journal of Medical Systems, 2010, 34(2): 195-212.

[27] Natarajan K, Acharya U R, Alias F, et al. Nonlinear analysis of EEG signals at different mental states[J]. Biomedical Engineering Online, 2004, 3(1): 1-11.

[28] Baldwin C L, Penaranda B N. Adaptive training using an artificial neural network and EEG metrics for within-and cross-task workload classification[J]. Neuroimage, 2012, 59(1): 48-56.

[29] Brouwer A M, Hogervorst M A, Van Erp J B F, et al. Estimating workload using EEG spectral power and ERPs in the n-back task[J]. Journal of Neural Engineering, 2012, 9(4): 045008.

[30] Wang Z, Hope R M, Wang Z, et al. Cross-subject workload classification with a hierarchical Bayes model[J]. Neuroimage, 2012, 59(1): 64-69.

[31] Oviatt S. Human-centered design meets cognitive load theory: designing interfaces that help people think[C]//Proceedings of the 14th ACM international conference on Multimedia. 2006: 871-880.

[32] Moray N. Models and measures of mental workload[J]. Mental workload: Its theory and measurement, Springer, Boston, MA.1979: 13-21.

[33] Young M S, Stanton N A. Mental workload[M]//Handbook of human factors and ergonomics methods. CRC Press, London: Taylor & Francis, 2004: 416-426.

[34] Hart S G, Wickens C D. Workload assessment and prediction[M]// Manprint: An approach to systems integration. Dordrecht: Springer Netherlands, 1990: 257–296.

[35] Wickens C. Multiple resources and performance prediction[J]. Theoretical Issues in Ergonomics Science, 2022, 3:159 - 177

[36] O’Donnell R D, Eggemeier F T. Workload assessment methodology[M]//Handbook of perception and human performance, Vol. 2: Cognitive processes and performance. Oxford, England: John Wiley & Sons, 1986: 1–49.

[37] Rouse W B, Edwards S L, Hammer J M. Modeling the dynamics of mental workload and human performance in complex systems[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1993, 23(6): 1662–1671.

[38] Sciaraffa N, Aricò P, Borghini G, et al. On the use of machine learning for EEG-based Workload assessment: Algorithms comparison in a realistic task[C]//Human Mental Workload: Models and Applications: Third International Symposium, H-WORKLOAD 2019, Rome, Italy, November 14–15, 2019, Proceedings 3. Springer International Publishing, 2019: 170-185.

[39] Dimitrakopoulos G N, Kakkos I, Dai Z, et al. Task-independent mental workload classification based upon common multiband EEG cortical connectivity[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25(11): 1940-1949.

[40] Müller K R, Tangermann M, Dornhege G, et al. Machine learning for real-time single-trial EEG-analysis: from brain–computer interfacing to mental state monitoring[J]. Journal of neuroscience methods, 2008, 167(1): 82-90.

[41] Wang S, Gwizdka J, Chaovalitwongse W A. Using wireless EEG signals to assess memory workload in the n-back task[J]. IEEE Transactions on Human-Machine Systems, 2015, 46(3): 424-435.

[42] Roy R N, Charbonnier S, Campagne A, et al. Efficient mental workload estimation using task-independent EEG features[J]. Journal of neural engineering, 2016, 13(2): 026019.

[43] Zarjam P, Epps J, Lovell N H. Beyond subjective self-rating: EEG signal classification of cognitive workload[J]. IEEE Transactions on Autonomous Mental Development, 2015, 7(4): 301-310.

[44] Friedman N, Fekete T, Gal K, et al. EEG-based prediction of cognitive load in intelligence tests[J]. Frontiers in human neuroscience, 2019, 13: 191.

[45] Knoll A, Wang Y, Chen F, et al. Measuring cognitive workload with low-cost electroencephalograph[C]//Human-Computer Interaction–INTERACT 2011: 13th IFIP TC 13 International Conference, Lisbon, Portugal, September 5-9, 2011, Proceedings, Part IV 13. Springer Berlin Heidelberg, 2011: 568-571.

[46] Yu K, Prasad I, Mir H, et al. Cognitive workload modulation through degraded visual stimuli: A single-trial EEG study[J]. Journal of neural engineering, 2015, 12(4): 046020.

[47] Arico P, Borghini G, Di Flumeri G, et al. Reliability over time of EEG-based mental workload evaluation during Air Traffic Management (ATM) tasks[C]//2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2015: 7242-7245.

[48] Almogbel M A, Dang A H, Kameyama W. Cognitive workload detection from raw EEG-signals of vehicle driver using deep learning[C]//2019 21st International Conference on Advanced Communication Technology (ICACT). IEEE, 2019: 1-6.

[49] Dehais F, Duprès A, Blum S, et al. Monitoring pilot’s mental workload using ERPs and spectral power with a six-dry-electrode EEG system in real flight conditions[J]. Sensors, 2019, 19(6): 1324.

[50] Fournier L R, Wilson G F, Swain C R. Electrophysiological, behavioral, and subjective indexes of workload when performing multiple tasks: manipulations of task difficulty and training[J]. International Journal of Psychophysiology, 1999, 31(2): 129-145.

[51] Prinzel III L J, Freeman F G, Scerbo M W, et al. Effects of a psychophysiological system for adaptive automation on performance, workload, and the event-related potential P300 component[J]. Human factors, 2003, 45(4): 601-614.

[52] 尹钟. 基于生理特征与支持向量机的认知任务负荷瞬时识别[D]. 华东理工大学, 2015.

[53] Yin Z, Zhao M, Zhang W, et al. Physiological-signal-based mental workload estimation via transfer dynamical autoencoders in a deep learning framework[J]. Neurocomputing, 2019, 347: 212-229.

[54] Antonenko P, Paas F, Grabner R, et al. Using electroencephalography to measure cognitive load[J]. Educational psychology review, 2010, 22(4): 425-438.

[55] Mazher M, Abd Aziz A, Malik A S, et al. An EEG-based cognitive load assessment in multimedia learning using feature extraction and partial directed coherence[J]. IEEE Access, 2017, 5: 14819-14829.

[56] Jimenez-Molina A, Retamal C, Lira H. Using psychophysiological sensors to assess mental workload during web browsing[J]. Sensors, 2018, 18(2): 458.

[57] Devlin S P, Brown N L, Drollinger S, et al. Scan-based eye tracking measures are predictive of workload transition performance[J]. Applied Ergonomics, 2022, 105: 103829.

[58] Fernandez Rojas R, Debie E, Fidock J, et al. Electroencephalographic workload indicators during teleoperation of an unmanned aerial vehicle shepherding a swarm of unmanned ground vehicles in contested environments[J]. Frontiers in neuroscience, 2020, 14: 40.

[59] Aricò P, Borghini G, Di Flumeri G, et al. A passive brain–computer interface application for the mental workload assessment on professional air traffic controllers during realistic air traffic control tasks[J]. Progress in brain research, 2016, 228: 295-328.

[60] 杨柳, 何萌, 刘清. 船员工作负荷检测方法研究综述[J]. 交通信息与安全, 2021, 39(3): 1-7, 16.

[61] Zhang L, Wade J, Bian D, et al. Cognitive load measurement in a virtual reality-based driving system for autism intervention[J]. IEEE transactions on affective computing, 2017, 8(2): 176-189.

[62] Yüce A, Gao H, Cuendet G L, et al. Action units and their cross-correlations for prediction of cognitive load during driving[J]. IEEE Transactions on Affective Computing, 2016, 8(2): 161-175.

[63] Cheng B, Fan C, Fu H, et al. Measuring and computing cognitive statuses of construction workers based on electroencephalogram: a critical review[J]. IEEE Transactions on Computational Social Systems, 2022.

[64] Abrantes A, Comitz E, Mosaly P, et al. Classification of EEG Features for Prediction of Working Memory Load[C]//Advances in The Human Side of Service Engineering: Proceedings of the AHFE 2016 International Conference on The Human Side of Service Engineering, July 27-31, 2016, Walt Disney World®, Florida, USA. Springer International Publishing, 2017: 115-126.

[65] Mathan S, Smart A, Ververs T, et al. Towards an index of cognitive efficacy EEG-based estimation of cognitive load among individuals experiencing cancer-related cognitive decline[C]//2010 Annual International Conference of the IEEE Engineering in Medicine and Biology. IEEE, 2010: 6595-6598.

[66] Wu W, Zhang Y, Jiang J, et al. An electroencephalographic signature predicts antidepressant response in major depression[J]. Nature biotechnology, 2020, 38(4): 439-447.

[67] Du X, Li J, Xiong D, et al. Research on electroencephalogram specifics in patients with schizophrenia under cognitive load[J]. Journal of Biomedical Engineering, 2020, 37(1): 45-53.

[68] De Waard D, Brookhuis K A. The measurement of drivers' mental workload[D]. University of Groningen, 1996.

[69] Cain B. A review of the mental workload literature[J]. Toronto: Defence Research and Development，2007.

[70] 柯余峰. 脑力负荷的脑电响应, 识别与自适应脑—机交互技术研究[D]. 天津大学, 2017.

[71] 卫宗敏，郝红勋，徐其志，等．飞行员脑力负荷测量指标和评价方法研究进展[J]．科学技术与工程，2019，19(24) : 1-8.

[72] Hart S G, Staveland L E. Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research[M]//Advances in psychology. North-Holland, 1988, 52: 139-183.

[73] Reid G B, Nygren T E. The subjective workload assessment technique: A scaling procedure for measuring mental workload[M]//Advances in psychology. North-Holland, 1988, 52: 185-218.

[74] Hicks T G, Wierwille W W. Comparison of five mental workload assessment procedures in a moving-base driving simulator[J]. Human factors, 1979, 21(2): 129-143.

[75] 王秋莲 主编. 人因工程[M]. 北京：科学出版社, 2022.

[76] Meshkati N, Loewenthal A. An eclectic and critical review of four primary mental workload assessment methods: A guide for developing a comprehensive model[J]. Advances in Psychology, 1988,52:251-267.

[77] Blankertz B, Tangermann M, KR Müller. Brain-Computer Interfaces: Principles and Practice[M]. 2012, P66.

[78] 科学技术部社会发展科技司，中国生物技术发展中心. 2020中国生命科学与生物技术发展报告[M].科学出版社，2020.

[79] Ramadan R A, Vasilakos A V. Brain computer interface: control signals review[J]. Neurocomputing, 2017, 223: 26-44.

[80] 沈敏. 脑-机接口技术综述[J]. 重庆邮电大学学报:自然科学版, 2007, 19(B06):147-150.

[81] 陈小刚, 王毅军, 张丹. 2018年脑机接口研发热点回眸[J]. 科技导报, 2019, 37(1): 173-179.

[82] 中国科学院创新发展研究中心,中国生命健康技术预见研究组. 中国生命健康2035技术预见[M].北京：科学出版社，2020.8.

[83] 明东. 脑-机接口:开启人机共融之路[J]. 民主与科学, 2017(2).

[84] Shanechi M M. Brain–machine interfaces from motor to mood[J]. Nature Neuroscience, 2019, 22(10):1554-1564.

[85] Wang P, Gong P, Zhou Y, et al. Decoding the continuous motion imagery trajectories of upper limb skeleton points for EEG-based brain-computer interface[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72:1-12.

[86] Hong X, Zheng Q, Liu L, et al. Dynamic joint domain adaptation network for motor imagery classification[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2021, 29: 556-565.

[87] Guger C, Da Ba N S, Sellers E, et al. How many people are able to control a P300-based brain-computer interface (BCI)? [J]. Neuroscience Letters, 2009, 462:94-98.

[88] Podmore J J, Breckon T P, Aznan N, et al. On the Relative Contribution of Deep Convolutional Neural Networks for SSVEP-Based Bio-Signal Decoding in BCI Speller Applications[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(4):611-618.

[89] Liu Y, Zhou Y, Zhang D. TcT: Temporal and channel Transformer for EEG-based Emotion Recognition[C]//2022 IEEE 35th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2022: 366-371.

[90] Gao L, Zhu L, Hu L, et al. Mid-task physical exercise keeps your mind vigilant: Evidences from behavioral performance and EEG functional connectivity[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2020, 29: 31-40.

[91] Mishchenko Y, Kaya M. Detecting the attention state of an operator in continuous attention task using EEG-based brain-computer interface[C]//2015 23nd Signal Processing and Communications Applications Conference (SIU). IEEE, 2015: 232-235.

[92] Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis[J]. Journal of neuroscience methods, 2004, 134(1): 9-21.

[93] Gramfort A, Luessi M, Larson E, et al. MEG and EEG data analysis with MNE-Python[J]. Frontiers in neuroscience, 2013, 7: 267.

[94] W. Peng.EEG Preprocessing and Denoising[M]//EEG Signal Processing and Feature Extraction, L. Hu and Z. Zhang, eds., Singapore: Springer Singapore, 2019, 71-87.

[95] Luck S J. An introduction to the event-related potential technique[M]. MIT press, 2014.

[96] Mognon A, Jovicich J, Bruzzone L, et al. ADJUST: An automatic EEG artifact detector based on the joint use of spatial and temporal features[J]. Psychophysiology, 2011, 48(2): 229-240.

[97] Mullen T, Kothe C, Chi Y M, et al. Real-time modeling and 3D visualization of source dynamics and connectivity using wearable EEG[C]//2013 35th annual international conference of the IEEE engineering in medicine and biology society (EMBC). IEEE, 2013: 2184-2187.

[98] Kok A. On the utility of P3 amplitude as a measure of processing capacity[J]. Psychophysiology, 2001, 38(3): 557-577.

[99] Stewart A X, Nuthmann A, Sanguinetti G. Single-trial classification of EEG in a visual object task using ICA and machine learning[J]. Journal of neuroscience methods, 2014, 228: 1-14.

[100] Kabbara A. Brain network estimation from dense EEG signals: application to neurological disorders. Neurons and Cognition, 2018.

[101] Mühl C, Jeunet C, Lotte F. EEG-based workload estimation across affective contexts[J]. Frontiers in neuroscience, 2014, 8: 114.

[102] Zarjam P, Epps J, Chen F. Characterizing working memory load using EEG delta activity[C]//2011 19th European Signal Processing Conference. IEEE, 2011: 1554-1558.

[103] Dijksterhuis C, De Waard D, Brookhuis K A, et al. Classifying visuomotor workload in a driving simulator using subject specific spatial brain patterns[J]. Frontiers in neuroscience, 2013, 7: 149.

[104] Zarjam P, Epps J, Chen F, et al. Estimating cognitive workload using wavelet entropy-based features during an arithmetic task[J]. Computers in biology and medicine, 2013, 43(12): 2186-2195.

[105] Borghini G, Vecchiato G, Toppi J, et al. Assessment of mental fatigue during car driving by using high resolution EEG activity and neurophysiologic indices[C]//2012 annual international conference of the IEEE engineering in medicine and biology society. IEEE, 2012: 6442-6445.

[106] Walter C, Schmidt S, Rosenstiel W, et al. Using cross-task classification for classifying workload levels in complex learning tasks[C]//2013 Humaine Association Conference on Affective Computing and Intelligent Interaction. IEEE, 2013: 876-881.

[107] Kıymık M K, Güler İ, Dizibüyük A, et al. Comparison of STFT and wavelet transform methods in determining epileptic seizure activity in EEG signals for real-time application[J]. Computers in biology and medicine, 2005, 35(7): 603-616.

[108] Subasi A. EEG signal classification using wavelet feature extraction and a mixture of expert model[J]. Expert Systems with Applications, 2007, 32(4): 1084-1093.

[109] Medl A. Time frequency and wavelets in biomedical signal processing[J]. IEEE Engineering in Medicine & Biology Magazine, 1998, 17(6): 15-97.

[110] Tian Y, Zhang H, Jiang Y, et al. A fusion feature for enhancing the performance of classification in working memory load with single-trial detection[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(10): 1985-1993.

[111] Roy R N, Bonnet S, Charbonnier S, et al. Mental fatigue and working memory load estimation: interaction and implications for EEG-based passive BCI[C]//2013 35th annual international conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2013: 6607-6610.

[112] Bashivan P, Yeasin M, Bidelman G M. Single trial prediction of normal and excessive cognitive load through EEG feature fusion[C]//2015 IEEE signal processing in medicine and biology symposium (SPMB). IEEE, 2015: 1-5.

[113] Appriou A, Cichocki A, Lotte F. Modern machine-learning algorithms: for classifying cognitive and affective states from electroencephalography signals[J]. IEEE Systems, Man, and Cybernetics Magazine, 2020, 6(3): 29-38.

[114] Ma Y, Shi W, Peng C K, et al. Nonlinear dynamical analysis of sleep electroencephalography using fractal and entropy approaches[J]. Sleep medicine reviews, 2018, 37: 85-93.

[115] Bai Y, Li X, and Liang Z. Nonlinear Neural Dynamics[M]// EEG Signal Processing and Feature Extraction, L. Hu and Z. Zhang, eds., Singapore: Springer Singapore, 2019, 215-240.

[116] Sakkalis V. Review of advanced techniques for the estimation of brain connectivity measured with EEG/MEG[J]. Computers in biology and medicine, 2011, 41(12): 1110-1117.

[117] Chakladar D D, Dey S, Roy P P, et al. EEG-based mental workload estimation using deep BLSTM-LSTM network and evolutionary algorithm[J]. Biomedical Signal Processing and Control, 2020, 60: 101989.

[118] Zhang J, Yin Z, Wang R. Pattern classification of instantaneous cognitive task-load through gmm clustering, laplacian eigenmap, and ensemble svms[J]. IEEE/ACM transactions on computational biology and bioinformatics, 2016, 14(4): 947-965.

[119] Yin Z, Zhang J. Operator functional state classification using least-square support vector machine based recursive feature elimination technique[J]. Computer methods and programs in biomedicine, 2014, 113(1): 101-115.

[120] Theodoridis S, Koutroumbas K. Pattern recognition[M]. Elsevier, 2006.

[121] Gevins A, Smith M E, Leong H, et al. Monitoring working memory load during computer-based tasks with EEG pattern recognition methods[J]. Human factors, 1998, 40(1): 79-91.

[122] Wilson G F, Russell C A. Operator functional state classification using multiple psychophysiological features in an air traffic control task[J]. Human Factors, 2003, 45(3): 381-389.

[123] Wilson G F, Russell C A. Real-time assessment of mental workload using psychophysiological measures and artificial neural networks[J]. Human factors, 2003, 45(4): 635-644.

[124] Wilson G F, Russell C A. Performance enhancement in an uninhabited air vehicle task using psychophysiologically determined adaptive aiding[J]. Human factors, 2007, 49(6): 1005-1018.

[125] Ye J, Janardan R, Li Q. Two-dimensional linear discriminant analysis[J]. Advances in neural information processing systems, 2004, 17.

[126] Kohlmorgen J, Dornhege G, Braun M, et al. Improving human performance in a real operating environment through real-time mental workload detection[J]. Toward brain-computer interfacing, 2007, 409422: 409-422.

[127] Tremmel C, Herff C, Sato T, et al. Estimating cognitive workload in an interactive virtual reality environment using EEG[J]. Frontiers in human neuroscience, 2019, 13: 401.

[128] Borghini G, Aricò P, Di Flumeri G, et al. EEG-based cognitive control behaviour assessment: an ecological study with professional air traffic controllers[J]. Scientific reports, 2017, 7(1): 547.

[129] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. nature, 2015, 521(7553): 436-444.

[130] Guo J, Lei Z, Wan J, et al. Dominant and complementary emotion recognition from still images of faces[J]. IEEE Access, 2018, 6: 26391-26403.

[131] Gao Z, Wang X, Yang Y, et al. EEG-based spatio–temporal convolutional neural network for driver fatigue evaluation[J]. IEEE transactions on neural networks and learning systems, 2019, 30(9): 2755-2763.

[132] Ma X, Qiu S, Du C, et al. Improving EEG-based motor imagery classification via spatial and temporal recurrent neural networks[C]//2018 40th annual international conference of the IEEE engineering in medicine and biology society (EMBC). IEEE, 2018: 1903-1906.

[133] Zheng W L, Zhu J Y, Peng Y, et al. EEG-based emotion classification using deep belief networks[C]//2014 IEEE international conference on multimedia and expo (ICME). IEEE, 2014: 1-6.

[134] Qiao W, Bi X. Ternary-task convolutional bidirectional neural turing machine for assessment of EEG-based cognitive workload[J]. Biomedical Signal Processing and Control, 2020, 57: 101745.

[135] Craik A, He Y, Contreras-Vidal J L. Deep learning for electroencephalogram (EEG) classification tasks: a review[J]. Journal of neural engineering, 2019, 16(3): 031001.

[136] Blanco J A, Johnson M K, Jaquess K J, et al. Quantifying cognitive workload in simulated flight using passive, dry EEG measurements[J]. IEEE Transactions on Cognitive and Developmental Systems, 2016, 10(2): 373-383.

[137] Kuanar S, Athitsos V, Pradhan N, et al. Cognitive analysis of working memory load from EEG, by a deep recurrent neural network[C]//2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2018: 2576-2580.

[138] Hefron R G, Borghetti B J, Christensen J C, et al. Deep long short-term memory structures model temporal dependencies improving cognitive workload estimation[J]. Pattern Recognition Letters, 2017, 94: 96-104.

[139] Zhang J, Li S. A deep learning scheme for mental workload classification based on restricted Boltzmann machines[J]. Cognition, Technology & Work, 2017, 19(4): 607-631.

[140] Yin Z, Zhang J. Cross-subject recognition of operator functional states via EEG and switching deep belief networks with adaptive weights[J]. Neurocomputing, 2017, 260(18): 349-366.

[141] Fan Y J. Autoencoder node saliency: Selecting relevant latent representations[J]. Pattern Recognition, 2019, 88: 643-653.

[142] Li J, Struzik Z, Zhang L, et al. Feature learning from incomplete EEG with denoising autoencoder[J]. Neurocomputing, 2015, 165: 23-31.

[143] Yin Z, Zhang J. Recognition of cognitive task load levels using single channel EEG and stacked denoising autoencoder[C]//2016 35th Chinese Control Conference (CCC). IEEE, 2016: 3907-3912.

[144] Yin Z, Zhang J. Cross-session classification of mental workload levels using EEG and an adaptive deep learning model[J]. Biomedical Signal Processing and Control, 2017, 33: 30-47.

[145] Yang S, Yin Z, Wang Y, et al. Assessing cognitive mental workload via EEG signals and an ensemble deep learning classifier based on denoising autoencoders[J]. Computers in biology and medicine, 2019, 109: 159-170.

[146] Ke Y, Qi H, Zhang L, et al. Towards an effective cross-task mental workload recognition model using electroencephalography based on feature selection and support vector machine regression[J]. International Journal of Psychophysiology, 2015, 98(2): 157–166.

[147] Kakkos I, Dimitrakopoulos G N, Sun Y, et al. EEG fingerprints of task-independent mental workload discrimination[J]. IEEE Journal of Biomedical and Health Informatics, 2021, 25(10): 3824-3833.

[148] Zhang P, Wang X, Zhang W, et al. Learning spatial–spectral–temporal EEG features with recurrent 3D convolutional neural networks for cross-task mental workload assessment[J]. IEEE Transactions on neural systems and rehabilitation engineering, 2018, 27(1): 31-42.

[149] Guan K, Zhang Z, Chai X, et al. EEG Based Dynamic Functional Connectivity Analysis in Mental Workload Tasks with Different Types of Information[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022, 30: 632-642.

[150] Gupta S S, Taori T J, Ladekar M Y, et al. Classification of cross task cognitive workload using deep recurrent network with modelling of temporal dynamics[J]. Biomedical Signal Processing and Control, 2021, 70: 103070.

[151] Gevins A, Smith M E, Leong H, et al. Monitoring working memory load during computer-based tasks with EEG pattern recognition methods[J]. Human factors, 1998, 40(1): 79-91.

[152] Bashivan P, Rish I, Yeasin M, et al. Learning representations from EEG with deep recurrent-convolutional neural networks[J]. arXiv preprint arXiv:1511.06448, 2015.

[153] Ni Z, Xu J, Wu Y, et al. Improving cross-state and cross-subject visual ERP-based BCI with temporal modeling and adversarial training[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022, 30: 369-379.

[154] Zhou Y, Huang S, Xu Z, et al. Cognitive workload recognition using EEG signals and machine learning: a review[J]. IEEE Transactions on Cognitive and Developmental Systems, 2021.

[155] Lan Z, Sourina O, Wang L, et al. Domain adaptation techniques for EEG-based emotion recognition: a comparative study on two public datasets[J]. IEEE Transactions on Cognitive and Developmental Systems, 2018, 11(1): 85-94.

[156] Wu D, Xu Y, Lu B L. Transfer learning for EEG-based brain–computer interfaces: A review of progress made since 2016[J]. IEEE Transactions on Cognitive and Developmental Systems, 2020, 14(1): 4-19.

[157] Zheng W L, Lu B L. Personalizing EEG-based affective models with transfer learning[C]//Proceedings of the twenty-fifth international joint conference on artificial intelligence. 2016: 2732-2738.

[158] Schneider W, Eschman A, Zuccolotto A. E-Prime Reference Guide. Pittsburgh, PA, USA: Psychology Software Tools, 2002.

[159] Pion-Tonachini L, Kreutz-Delgado K, Makeig S. ICLabel: An automated electroencephalographic independent component classifier, dataset, and website[J]. NeuroImage, 2019, 198: 181-197.

[160] Zimmerman D W, Zumbo B D. Relative power of the Wilcoxon test, the Friedman test, and repeated-measures ANOVA on ranks[J]. The Journal of Experimental Education, 1993, 62(1): 75-86.

[161] Dinno A. Nonparametric pairwise multiple comparisons in independent groups using Dunn's test[J]. The Stata Journal, 2015, 15(1): 292-300.

[162] Makeig S. Auditory event-related dynamics of the EEG spectrum and effects of exposure to tones[J]. Electroencephalography and clinical neurophysiology, 1993, 86(4): 283-293.

[163] Yi W, Qiu S, Wang K, et al. Enhancing performance of a motor imagery based brain–computer interface by incorporating electrical stimulation-induced SSSEP[J]. Journal of neural engineering, 2017, 14(2): 026002.

[164] Pan S J, Tsang I W, Kwok J T, et al. Domain adaptation via transfer component analysis[J]. IEEE transactions on neural networks, 2010, 22(2): 199-210.

[165] Gretton A, Borgwardt K, Rasch M, et al. A kernel method for the two-sample-problem[C]//In Advances in neural information processing systems, 2006, 19.

[166] Long M, Wang J, Ding G, et al. Transfer feature learning with joint distribution adaptation[C]//Proceedings of the IEEE international conference on computer vision. 2013: 2200-2207.

[167] Wang J, Chen Y, Hao S, et al. Balanced distribution adaptation for transfer learning[C]//2017 IEEE international conference on data mining (ICDM). IEEE, 2017: 1129-1134.

[168] Long M, Wang J, Ding G, et al. Transfer joint matching for unsupervised domain adaptation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 1410-1417.

[169] He B, Astolfi L, Valdés-Sosa P A, et al. Electrophysiological brain connectivity: theory and implementation[J]. IEEE transactions on biomedical engineering, 2019, 66(7): 2115-2137.

[170] Zhou Y, Qiao L, Li W, et al. Simultaneous estimation of low-and high-order functional connectivity for identifying mild cognitive impairment[J]. Frontiers in neuroinformatics, 2018, 12: 3.

[171] Zhou Y, Zhang L, Teng S, et al. Improving sparsity and modularity of high-order functional connectivity networks for MCI and ASD identification[J]. Frontiers in neuroscience, 2018, 12: 959.

[172] Lachaux J P, Lutz A, Rudrauf D, et al. Estimating the time-course of coherence between single-trial brain signals: an introduction to wavelet coherence[J]. Neurophysiology Clinique, 2002, 32(3): 157-174.

[173] Laurens V D M, Hinton G. Visualizing data using t-SNE[J]. Journal of machine learning research, 2008, 9(11): 2579-2605.

[174] Li R, Johansen J S, Ahmed H, et al. The perils and pitfalls of block design for EEG classification experiments[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 43(1): 316-333.

[175] Wan Z, Yang R, Huang M, et al. A review on transfer learning in EEG signal analysis[J]. Neurocomputing, 2021, 421: 1-14.

[176] Wei W, Qiu S, Ma X, et al. Reducing calibration efforts in RSVP tasks with multi-source adversarial domain adaptation[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2020, 28(11): 2344-2355.

[177] Wei C S, Lin Y P, Wang Y T, et al. A subject-transfer framework for obviating inter-and intra-subject variability in EEG-based drowsiness detection[J]. Neuroimage, 2018, 174: 407-419.

[178] Zhang P, Wang X, Chen J, et al. Spectral and temporal feature learning with two-stream neural networks for mental workload assessment[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(6): 1149-1159.

[179] Gupta A, Siddhad G, Pandey V, et al. Subject-specific cognitive workload classification using EEG-based functional connectivity and deep learning[J]. Sensors, 2021, 21(20): 6710.

[180] Liu Y, Lan Z, Cui J, et al. Inter-subject transfer learning for EEG-based mental fatigue recognition[J]. Advanced Engineering Informatics, 2020, 46: 101157.

[181] Zhang J, Wang Y, Li S. Cross-subject mental workload classification using kernel spectral regression and transfer learning techniques[J]. Cognition, Technology & Work, 2017, 19(4): 587-605.

[182] Bhosale S, Chakraborty R, Kopparapu S K. Calibration free meta learning based approach for subject independent EEG emotion recognition[J]. Biomedical Signal Processing and Control, 2022, 72: 103289.

[183] Plechawska-Wójcik M, Tokovarov M, Kaczorowska M, et al. A three-class classification of cognitive workload based on EEG spectral data[J]. Applied Sciences, 2019, 9(24): 5340.

[184] Appriou A, Cichocki A, Lotte F. Towards robust neuroadaptive HCI: exploring modern machine learning methods to estimate mental workload from EEG signals[C]//Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems. 2018: 1-6.

[185] Jiao Z, Gao X, Wang Y, et al. Deep convolutional neural networks for mental load classification based on EEG data[J]. Pattern Recognition, 2018, 76: 582-595.

[186] Hefron R, Borghetti B, Schubert Kabban C, et al. Cross-participant EEG-based assessment of cognitive workload using multi-path convolutional recurrent neural networks[J]. Sensors, 2018, 18(5): 1339.

[187] So W K Y, Wong S W H, Mak J N, et al. An evaluation of mental workload with frontal EEG[J]. PloS one, 2017, 12(4): e0174949.

[188] Zhuang F, Qi Z, Duan K, et al. A comprehensive survey on transfer learning[J]. Proceedings of the IEEE, 2020, 109(1): 43-76.

[189] Weiss K, Khoshgoftaar T M, Wang D D. A survey of transfer learning[J]. Journal of Big data, 2016, 3(1): 1-40.

[190] Li Y, Zheng W, Zong Y, et al. A bi-hemisphere domain adversarial neural network model for EEG emotion recognition[J]. IEEE Transactions on Affective Computing, 2018, 12(2): 494-504.

[191] Latif S, Rana R, Khalifa S, et al. Self-supervised adversarial domain adaptation for cross-corpus and cross-language speech emotion recognition[J]. IEEE Transactions on Affective Computing, 2022.

[192] Jiménez-Guarneros M, Gómez-Gil P. Custom Domain Adaptation: A new method for cross-subject, EEG-based cognitive load recognition[J]. IEEE Signal Processing Letters, 2020, 27: 750-754.

[193] Long M, Cao Y, Cao Z, et al. Transferable representation learning with deep adaptation networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2018, 41(12): 3071-3085.

[194] Li J, Qiu S, Du C, et al. Domain adaptation for EEG emotion recognition based on latent representation similarity[J]. IEEE Transactions on Cognitive and Developmental Systems, 2019, 12(2): 344-353.

[195] Ganin Y, Ustinova E, Ajakan H, et al. Domain-adversarial training of neural networks[J]. The journal of machine learning research, 2016, 17(1): 2096-2030.

[196] Zhao H, Zheng Q, Ma K, et al. Deep representation-based domain adaptation for nonstationary EEG classification[J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 32(2): 535-545.

[197] https://www.neuroergonomicsconference.um.ifi.lmu.de/pbci/.

[198] Lawhern V J, Solon A J, Waytowich N R, et al. EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces[J]. Journal of neural engineering, 2018, 15(5): 056013.

[199] Schirrmeister R T, Springenberg J T, Fiederer L D J, et al. Deep learning with convolutional neural networks for EEG decoding and visualization[J]. Human brain mapping, 2017, 38(11): 5391-5420.

[200] Long M, Cao Z, Wang J, et al. Conditional adversarial domain adaptation[J]. Advances in neural information processing systems, 2018, 31: 1640–1650.

[201] Abadi M, Agarwal A, Barham P, et al. Tensorflow: Large-scale machine learning on heterogeneous distributed systems[J]. arXiv preprint arXiv:1603.04467, 2016.

[202] Kingma D P, Ba J. Adam: A method for stochastic optimization[J]. arXiv preprint arXiv:1412.6980, 2014.

[203] Woolson R F. Wilcoxon signed‐rank test[J]. Wiley encyclopedia of clinical trials, 2007: 1-3.

[204] Long M, Zhu H, Wang J, et al. Deep transfer learning with joint adaptation networks[C]//International conference on machine learning. PMLR, 2017: 2208-2217.

[205] Li S, Liu C H, Xie B, et al. Joint adversarial domain adaptation[C]//Proceedings of the 27th ACM International Conference on Multimedia. New York, NY, USA, 2019: 729-737.

[206] Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. Salt Lake City, UT, USA, 2018: 7132-7141.

[207] Saito K, Watanabe K, Ushiku Y, et al. Maximum classifier discrepancy for unsupervised domain adaptation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. Salt Lake City, UT, USA,2018: 3723-3732.

[208] Du Z, Li J, Su H, et al. Cross-domain gradient discrepancy minimization for unsupervised domain adaptation[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021: 3937-3946.

[209] Wen Y, Zhang K, Li Z, et al. A discriminative feature learning approach for deep face recognition[C]//Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part VII 14. Springer International Publishing, 2016: 499-515.

[210] Zhou K, Liu Z, Qiao Y, et al. Domain generalization: A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022.

[211] Ingolfsson T. M., Hersche.M, Wang X, et al. EEG-TCNet: an accurate temporal convolutional network for embedded motor-imagery brain–machine interfaces[C]. IEEE International Conference on Systems, Man, and Cybernetics (SMC), Toronto, ON, Canada, 2020, 2958-2965.

[212] Roy RN, Hinss MF, Darmet L, et al. Retrospective on the First Passive Brain-Computer Interface Competition on Cross-Session Workload Estimation[J]. Frontiers in Neuroergonomics, 2022, 3:838342.

[213] Congedo M., Barachant A., and Bhatia R. Riemannian geometry for EEG-based brain-computer interfaces; a primer and a review[J]. Brain Comput. Interfaces, 2017, 4: 155–174.

[214] Tour DL, Moreau T, Jas M, et al. Multivariate convolutional sparse coding for electromagnetic brain signals[C]. In Proceedings of the 32nd International Conference on Neural Information Processing Systems (NIPS'18). Curran Associates Inc., Red Hook, NY, USA, 2018, 3296–3306.

[215] Kim H, Yoon J, Jeong B, et al. Rank-1 convolutional neural network[J]. arXiv preprint arXiv:1808.04303, 2018.