With the rapid development of intelligent technology, the gradual popularization of digital learning has provided learners with diversed resources and pathways to learn. However, while this learning mode enriches the learning experience, it also brings complex challenges to learners' cognitive load. Traditional cognitive load measurement methods have problems with accuracy, such as insufficient processing, a lack of dynamic monitoring, and a lack of immediate feedback. Therefore, intelligent assessment of cognitive load has gradually become a popular research topic. Intelligent assessment can achieve accurate assessment of cognitive load by combining physiological indicators and using non-invasive technical means such as machine learning and deep learning. This study systematically reviews the applications of traditional machine learning algorithms, such as Support Vector Machine (SVM), Random Forest, and Linear Discriminant Analysis (LDA), as well as deep learning algorithms, including Feedforward Neural Networks (FNN), Convolutional Neural Networks (CNN), and Long Short-Term Memory networks (LSTM), in the intelligent assessment of cognitive load. It also explores the applications of the hybrid models. Finally, the study proposes insights and development directions for the intelligent assessment of cognitive load in digital learning and explores the potential of the large language model(LLM) in the intelligent assessment of cognitive load, aiming to provide assistance for efficient digital learning.
在深度学习领域,卷积神经网络(Convolutional Neural Network,CNN)因其在图像处理任务中的卓越性能而受到广泛关注[52]。CNN的设计初衷是模仿生物视觉皮层的处理机制,通过局部感受野和权重共享来提取图像特征,随着研究的深入,CNN的变体被广泛应用于计算机视觉、自然语言处理(Natural Language Processing,NLP)、图像分割、遥感以及信号处理等多个领域[53-55]。CNN模型通过分层方式从数据样本中提取特征,其中卷积层负责学习复杂特征,池化层在提升模型性能的同时降低特征图维度,全连接层则将复杂特征映射至输出端[56]。在训练过程中,CNN不断优化权重及其他参数,尽管这一过程耗时,但训练完成后的分类速度会显著提升[57]。在认知负荷测评领域的应用中,CNN模型的输入策略、特征提取和选择方法随着架构的不同而变化[58]。
近年来,以Transformer架构为代表的预训练大语言模型(Large Language Model,LLM)已成为深度学习技术的集大成者,其通过海量数据的自监督预训练,展现出强大的特征表征能力与跨任务迁移适应性。LLM的核心优势在于其能够从高维、非线性、时序性数据中自动提取深层语义特征,并通过微调(Fine-tuning)策略快速适配下游任务,这一特性为认知负荷智能化测评中生理信号的特征提取与模式识别提供了新的技术路径[82]。尤其在面对EEG、ECG、ERP等生理信号的复杂时空特性时,传统机器学习方法常受限于手动特征提取的局限性,而LLM通过端到端的自注意力机制与层次化特征抽象,有望突破现有技术瓶颈,实现更精准的认知负荷动态监测与评估。
SWELLERJ. Cognitive load during problem solving: Effects on learning[J]. Cognitive Science, 1988, 12(2): 257-285. DOI:10.1016/0364-0213(88)90023-7 .
[2]
SWELLERJ, VAN MERRIËNBOERJ J G, PAASF. Cognitive architecture and instructional design: 20 years later[J]. Educational Psychology Review, 2019, 31(2): 261-292. DOI:10.1007/s10648-019-09465-5 .
[3]
SWELLERJ. Cognitive load theory and educational technology[J]. Educational Technology Research and Development, 2020, 68(1): 1-16. DOI:10.1007/s11423-019-09701-3 .
[4]
JEONGE J, BIOCCAF A. Are there optimal levels of arousal to memory? Effects of arousal, centrality, and familiarity on brand memory in video games[J]. Computers in Human Behavior, 2012, 28(2): 285-291. DOI:10.1016/j.chb.2011.09.011 .
[5]
JORDANJ, WAGNERJ, MANTHEYD E, et al. Optimizing lectures from a cognitive load perspective[J]. AEM Education and Training, 2019, 4(3): 306-312. DOI:10.1002/aet2.10389 .
[6]
SKULMOWSKIA, XUK M. Understanding cognitive load in digital and online learning: A new perspective on extraneous cognitive load[J]. Educational Psychology Review, 2022, 34(1): 171-196. DOI:10.1007/s10648-021-09624-7 .
[7]
SEPPS, WONGM, HOOGERHEIDEV, et al. Shifting online: 12 tips for online teaching derived from contemporary educational psychology research[J]. Journal of Computer Assisted Learning, 2022, 38(5): 1304-1320. DOI:10.1111/jcal.12715 .
[8]
KORBACHA, BRÜNKENR, PARKB. Differentiating different types of cognitive load: A comparison of different measures[J]. Educational Psychology Review, 2018, 30(2): 503-529. DOI:10.1007/s10648-017-9404-8 .
[9]
SWELLERJ. Element interactivity and intrinsic, extraneous, and germane cognitive load[J]. Educational Psychology Review, 2010, 22(2): 123-138. DOI:10.1007/s10648-010-9128-5 .
MANS J, JIANGQ, LIQ, et al. How CSCL affects digital learning resilience of college students—Analysis of mediating effects of cognitive load[J]. Modern Distance Education, 2023(1): 71-80. DOI: 10.13927/j.cnki.yuan.20230317.002 (Ch ).
GAOJ, PENGS D. Contradictions and strategies: Learning analytics based on extrinsic cognitive load in digital learning[J]. e-Education Research, 2023, 44(1): 100-105. DOI: 10.13811/j.cnki.eer.2023.01.014(Ch ).
[14]
PAASF G W C. Training strategies for attaining transfer of problem-solving skill in statistics: A cognitive-load approach[J]. Journal of Educational Psychology, 1992, 84(4): 429-434. DOI:10.1037/0022-0663.84.4.429 .
[15]
LEPPINKJ, PAASF, VAN DER VLEUTENC P M, et al. Development of an instrument for measuring different types of cognitive load[J]. Behavior Research Methods, 2013, 45(4): 1058-1072. DOI:10.3758/s13428-013-0334-1 .
[16]
REIDG B, NYGRENT E. The subjective workload assessment technique: A scaling procedure for measuring mental workload[J]. Advances in Psychology,1988,52: 185-218. DOI:10.1016/S0166-4115(08)62387-0 .
[17]
HARTS G, STAVELANDL E. Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research[J]. Advances in Psychology, 1988,52: 139-183. DOI:10.1016/S0166-4115(08)62386-9 .
[18]
WANGJ H, ANTONENKOP, KEILA, et al. Converging subjective and psychophysiological measures of cognitive load to study the effects of instructor-present video[J]. Mind, Brain, and Education, 2020, 14(3): 279-291. DOI:10.1111/mbe.12239 .
[19]
JIAS W, QIC C, CHENL L, et al. The effect of working memory load on feedback processing: Evidence from an event-related potentials (ERP) study[J]. Acta Psychologica Sinica, 2022, 54(3): 248-258. DOI:10.3724/sp.j.1041.2022.00248 .
[20]
VANGOG T, KESTERL, NIEVELSTEINF, et al. Uncovering cognitive processes: Different techniques that can contribute to cognitive load research and instruction[J]. Computers in Human Behavior, 2009, 25(2): 325-331. DOI:10.1016/j.chb.2008.12.021 .
[21]
NOURBAKHSHN, CHENF, WANGY, et al. Detecting users’ cognitive load by galvanic skin response with affective interference[J]. ACM Transactions on Interactive Intelligent Systems, 2017, 7(3): 1-20. DOI:10.1145/2960413 .
[22]
LEES, SHINH, HAHMC. Effective PPG sensor placement for reflected red and green light, and infrared wristband-type photoplethysmography[C]//2016 18th International Conference on Advanced Communication Technology (ICACT). New York: IEEE Press,2016: 556-558. DOI:10.1109/ICACT.2016.7423470 .
[23]
BROUWERA M, HOGERVORSTM A, VANERP 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. DOI:10.1088/1741-2560/9/4/045008 .
[24]
GEVINSA, SMITHM E, LEONGH, et al. Monitoring working memory load during computer-based tasks with EEG pattern recognition methods[J]. Human Factors, 1998, 40(1): 79-91. DOI:10.1518/001872098779480578 .
[25]
STENROOSM, SARVASJ. Bioelectromagnetic forward problem: Isolated source approach revis(it)ed[J]. Physics in Medicine and Biology, 2012, 57(11): 3517-3535. DOI:10.1088/0031-9155/57/11/3517 .
[26]
MOLTENIE, BASELLIG, CERUTTIS, et al. Load-dependent brain activation assessed by time-domain functional near-infrared spectroscopy during a working memory task with graded levels of difficulty[J]. Journal of Biomedical Optics, 2012, 17(5): 056005. DOI:10.1117/1.jbo.17.5.056005 .
[27]
HYÖNÄJ, TOMMOLAJ, ALAJAA M. Pupil dilation as a measure of processing load in simultaneous interpretation and other language tasks[J]. The Quarterly Journal of Experimental Psychology A, Human Experimental Psychology, 1995, 48(3): 598-612. DOI:10.1080/14640749508401407 .
[28]
GASPARJ, CARNEYC. The effect of partial automation on driver attention: A naturalistic driving study[J]. Human Factors, 2019, 61(8): 1261-1276. DOI:10.1177/0018720819836310 .
[29]
SIEGLEG J, ICHIKAWAN, STEINHAUERS. Blink before and after you think: Blinks occur prior to and following cognitive load indexed by pupillary responses[J]. Psychophysiology, 2008, 45(5): 679-687. DOI:10.1111/j.1469-8986.2008.00681.x .
[30]
HUGHESA M, HANCOCKG M, MARLOWS L, et al. Cardiac measures of cognitive workload: A meta-analysis[J]. Human Factors, 2019, 61(3): 393-414. DOI:10.1177/0018720819830553 .
[31]
CHAOC J, WUS Y, YAUY J, et al. Effects of three-dimensional virtual reality and traditional training methods on mental workload and training performance[J]. Human Factors and Ergonomics in Manufacturing & Service Industries, 2017, 27(4): 187-196. DOI:10.1002/hfm.20702 .
XUEY F, WANGK, QIUY S, et al. Cognitive load assessment of online learning based on multimodal data[J]. Modern Educational Technology, 2024, 34(3): 79-88 (Ch).
WANGG H, NIES X, XUER X. Cognitive load in multimedia learning: An overview of measurement methods and technologies[J]. e-Education Research, 2022, 43(4): 19-25. DOI: 10.13811/j.cnki.eer.2022.04.003 (Ch ).
[38]
ROYR N, CHARBONNIERS, CAMPAGNEA, et al. Efficient mental workload estimation using task-independent EEG features[J]. Journal of Neural Engineering, 2016, 13(2): 026019. DOI:10.1088/1741-2560/13/2/026019 .
[39]
PAHORA, SEITZA R, JAEGGIS M. Near transfer to an unrelated N-back task mediates the effect of N-back working memory training on matrix reasoning[J]. Nature Human Behaviour, 2022, 6(9): 1243-1256. DOI:10.1038/s41562-022-01384-w .
[40]
CHAKLADAR DDAS, ROYP P, IWAMURAM. EEG-based cognitive state classification and analysis of brain dynamics using deep ensemble model and graphical brain network[J]. IEEE Transactions on Cognitive and Developmental Systems, 2022, 14(4): 1507-1519. DOI:10.1109/TCDS.2021.3116079 .
[41]
ŞIMŞEKB, DIREKCIB, KOPARANB, et al. Examining the effect of augmented reality experience duration on reading comprehension and cognitive load[J]. Education and Information Technologies, 2025, 30(2): 1445-1464. DOI:10.1007/s10639-024-12864-z .
[42]
HOFFMANNJ A, VON HELVERSENB, RIESKAMPJ. Deliberation’s blindsight: How cognitive load can improve judgments [J]. Psychological Science, 2013, 24(6): 869-879. DOI:10.1177/0956797612463581 .
[43]
SAVAGES W, POTTERD D, TATLERB W. The effects of array structure and secondary cognitive task demand on processes of visual search[J]. Vision Research, 2018, 153: 37-46. DOI:10.1016/j.visres.2018.09.004 .
[44]
ZARJAMP, EPPSJ, LOVELLN H. Beyond subjective self-rating: EEG signal classification of cognitive workload[J]. IEEE Transactions on Autonomous Mental Development, 2015, 7(4): 301-310. DOI:10.1109/TAMD.2015.2441960 .
[45]
RAMAKRISHNANP, BALASINGAMB, BIONDIF. Cognitive load estimation for adaptive human-machine system automation[M]//Learning Control. Amsterdam: Elsevier,2021: 35-58. DOI:10.1016/B978-0-12-822314-7.00007-9 .
[46]
QUH Q, GAOX Y, PANGL P. Classification of mental workload based on multiple features of ECG signals[J]. Informatics in Medicine Unlocked, 2021, 24: 100575. DOI:10.1016/j.imu.2021.100575 .
MAY L, LUJ, ZHUJ Y, et al. Take-over performance prediction under different cognitive loads of non-driving tasks in highly automated driving[J]. Automotive Engineering, 2023, 45(12): 2330-2337. DOI: 10.19562/j.chinasae.qcgc.2023.12.015 (Ch ).
MENGR H, LIZ Q, WUZ J, et al. Study on cognitive load of operators based on assembly task and EEG power information features[J]. Journal of Machine Design, 2022, 39(11): 60-70. DOI: 10.13841/j.cnki.jxsj.2022.11.021(Ch ).
[53]
BISHOPC M. Pattern recognition and machine learning[M/OL]. New York: Springer, 2006. DOI: 10.1007/978-0-387-45528-0 .
[54]
ÇAKIRM P, VURALM, KOÇS Ö, et al. Real-time monitoring of cognitive workload of airline pilots in a flight simulator with fNIR optical brain imaging technology[C]// Foundations of Augmented Cognition: Neuroergonomics and Operational Neuroscience(9743). Cham: Springer International Publishing, 2016: 147-158. DOI:10.1007/978-3-319-39955-3_14 .
[55]
ABIBULLAEVB, ANJ. Classification of frontal cortex haemodynamic responses during cognitive tasks using wavelet transforms and machine learning algorithms[J]. Medical Engineering & Physics, 2012, 34(10): 1394-1410. DOI:10.1016/j.medengphy.2012.01.002 .
[56]
ZHAOS G, JIAOT H, ADADES Y S, et al. Based on vis-NIR combined with ANN for on-line detection of bacterial concentration during kombucha fermentation[J]. Food Bioscience, 2024, 60: 104346. DOI:10.1016/j.fbio.2024.104346 .
ZHAOX A, LIY, DONGP, et al. Artificial neural network models for local equilibrium scour depth around a single pile under wave or current actions[J/OL]. The Ocean Engineering, 2024: 1-18. (2024-12-23). DOI: 10.1016/j.oceaneng.2025.121102 .
[59]
TJOLLENGA, JUNGK, HONGW, et al. Classification of a Driver’s cognitive workload levels using artificial neural network on ECG signals[J]. Applied Ergonomics, 2017, 59: 326-332. DOI:10.1016/j.apergo.2016.09.013 .
[60]
SAMIMAS, SARMAM. Mental workload level assessment based on compounded hysteresis effect[J]. Cognitive Neurodynamics, 2023, 17(2): 357-372. DOI:10.1007/s11571-022-09830-1 .
PENGS Z, FANGZ J, GAOY B, et al. Crowd counting based on feature fusion of multi-scale fully convolutional networks[J]. Journal of Wuhan University (Natural Science Edition), 2018, 64(3): 249-254. DOI: 10.14188/j.1671-8836.2018.03.008 (Ch ).
[63]
BALASUNDARAMA, DILIPG, ASHOKKUMARS, et al. Detecting true medicinal leaves among similar leaves using computer vision and CNN[J]. National Academy Science Letters, 2023, 46(3): 257-262. DOI:10.1007/s40009-023-01210-5 .
[64]
BOULILAW, GHANDORHH, KHANM A, et al. A novel CNN-LSTM-based approach to predict urban expansion[J]. Ecological Informatics, 2021, 64: 101325. DOI:10.1016/j.ecoinf.2021.101325 .
[65]
AHMEDZ, WANGJ X. A fine-grained deep learning model using embedded-CNN with BiLSTM for exploiting product sentiments[J]. Alexandria Engineering Journal, 2023, 65: 731-747. DOI:10.1016/j.aej.2022.10.037 .
[66]
VENKATESWARAND, CHOY. Efficient solar power generation forecasting for greenhouses: A hybrid deep learning approach[J]. Alexandria Engineering Journal, 2024, 91: 222-236. DOI:10.1016/j.aej.2024.02.004 .
[67]
JEONGS, KIMB, CHA S, et al. Real-time CNN training and compression for neural-enhanced adaptive live streaming[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(9): 6023-6039. DOI:10.1109/TPAMI.2024.3377372 .
[68]
KHANM A, ASADIH, ZHANGL, et al. Application of artificial intelligence in cognitive load analysis using functional near-infrared spectroscopy: A systematic review[J]. Expert Systems with Applications, 2024, 249: 123717. DOI:10.1016/j.eswa.2024.123717 .
[69]
WANGL M, HUANGY H, CHOUP H, et al. Characteristics of brain connectivity during verbal fluency test: Convolutional neural network for functional near-infrared spectroscopy analysis[J]. Journal of Biophotonics, 2022, 15(1): e202100180. DOI:10.1002/jbio.202100180 .
[70]
KWONJ, IMC H. Subject-independent functional near-infrared spectroscopy-based brain-computer interfaces based on convolutional neural networks[J]. Frontiers in Human Neuroscience, 2021, 15: 646915. DOI:10.3389/fnhum.2021.646915 .
[71]
QINGK Q, HUANGR S, HONGK S. Decoding three different preference levels of consumers using convolutional neural network: A functional near-infrared spectroscopy study[J]. Frontiers in Human Neuroscience, 2021, 14: 597864. DOI:10.3389/fnhum.2020.597864 .
[72]
MILESG, SMITHM, ZOOKN, et al. EM-COGLOAD: An investigation into age and cognitive load detection using eye tracking and deep learning[J]. Computational and Structural Biotechnology Journal, 2024, 24: 264-280. DOI:10.1016/j.csbj.2024.03.014 .
[73]
ZHANGP B, WANGX, ZHANGW H, 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, 2019, 27(1): 31-42. DOI:10.1109/TNSRE.2018.2884641 .
[74]
CASCIANELLIS, BELLO-CEREZOR, BIANCONIF, et al. Dimensionality reduction strategies for CNN-based classification of histopathological images[M]//Intelligent Interactive Multimedia Systems and Services 2017. Cham: Springer International Publishing, 2017: 21-30. DOI:10.1007/978-3-319-59480-4_3 .
[75]
MAH J, LIUY L, RENY H, et al. Improved CNN classification method for groups of buildings damaged by earthquake, based on high resolution remote sensing images[J]. Remote Sensing, 2020, 12(2): 260. DOI:10.3390/rs12020260 .
XUC, LEIJ T, CHENY. A mechanism of proof-of-dynamic-trust based on LSTM and blacklist[J]. Journal of Wuhan University (Natural Science Edition), 2023, 69(2): 156-168. DOI: 10.14188/j.1671-8836.2022.0189 (Ch ).
[78]
ASGHERU, KHALILK, KHANM J, et al. Enhanced accuracy for multiclass mental workload detection using long short-term memory for brain-computer interface[J]. Frontiers in Neuroscience, 2020, 14: 584. DOI:10.3389/fnins.2020.00584 .
[79]
WANGQ, SMYTHED, CAOJ, et al. Characterisation of cognitive load using machine learning classifiers of electroencephalogram data[J]. Sensors, 2023, 23(20): 8528. DOI:10.3390/s23208528 .
[80]
YEDUKONDALUJ, SHARMAD, SHARMAL D. Subject-wise cognitive load detection using time-frequency EEG and Bi-LSTM[J]. Arabian Journal for Science and Engineering, 2024, 49(3): 4445-4457. DOI:10.1007/s13369-023-08494-1 .
[81]
ZHOUW G, YUP P, WUL H, et al. Pilot turning behavior cognitive load analysis in simulated flight[J]. Frontiers in Neuroscience, 2024, 18: 1450416. DOI:10.3389/fnins.2024.1450416 .
[82]
YOOG, KIMH, HONGS. Prediction of cognitive load from electroencephalography signals using long short-term memory network[J]. Bioengineering, 2023, 10(3): 361. DOI:10.3390/bioengineering10030361 .
ZHENGR L, ZHANGJ S. Assessing cognitive load combining features of time, frequency and spatial domain under digital graphical interface[J]. Journal of Computer-Aided Design & Computer Graphics, 2020, 32(7): 1062-1069. DOI: 10.3724/SP.J.1089.2020.18358.z54(Ch ).
[85]
RAHMANH, AHMEDM U, BARUAS, et al. Vision-based driver’s cognitive load classification considering eye movement using machine learning and deep learning[J]. Sensors, 2021, 21(23): 8019. DOI:10.3390/s21238019 .
[86]
CHAKLADAR DDAS, DEYS, ROYP 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. DOI:10.1016/j.bspc.2020.101989 .
[87]
CHAKLADAR DDAS, SAMANTAD, PRATIM ROYP. Multimodal deep sparse subspace clustering for multiple stimuli-based cognitive task[C]//2022 26th International Conference on Pattern Recognition (ICPR), New York: IEEE Press,2022: 1098-1104. DOI:10.1109/ICPR56361.2022.9955632 .
BANR Y, ZHOUD P, HANJ P, et al. Cognitive load recognition model based on improved attention mechanism in domain-adversarial networks[J]. Journal of Chinese Computer Systems, 2024, 45(11): 2602-2608. DOI: 10.20009/j.cnki.21-1106/TP.2024-0390 (Ch ).
[90]
KARPOUZIAN-ROGERST, SWEENEYJ A, RUBINL H, et al. Reduced task-evoked pupillary response in preparation for an executive cognitive control response among individuals across the psychosis spectrum[J]. Schizophrenia Research, 2022, 248: 79-88. DOI:10.1016/j.schres.2022.07.018 .
[91]
SUPRATAKA, DONGH, WUC, et al. DeepSleepNet: A model for automatic sleep stage scoring based on raw single-channel EEG[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25(11): 1998-2008. DOI:10.1109/TNSRE.2017.2721116 .
DAZAR, GOMEZL F, FIERREZJ, et al. DeepFace-attention: Multimodal face biometrics for attention estimation with application to e-learning[J]. IEEE Access, 2024, 12: 111343-111359. DOI:10.1109/ACCESS.2024.3437291 .
[94]
JIAOZ C, GAOX B, WANGY, et al. Deep Convolutional Neural Networks for mental load classification based on EEG data[J]. Pattern Recognition, 2018, 76: 582-595. DOI:10.1016/j.patcog.2017.12.002 .
[95]
WANGC, LIUY Y, GUOT Z, et al. Systems engineering issues for industry applications of large language model[J]. Applied Soft Computing, 2024, 151: 111165. DOI:10.1016/j.asoc.2023.111165 .
[96]
WANGG, LIUW, HEY, et al. EEGPT: Pretrained transformer for universal and reliable representation of EEG signals[J]. Advances in Neural Information Processing Systems, 2024, 37: 39249-39280.
ZHOUT T, ZHENGY Q, WEIT, et al. Cognitive load EEG recognition model integrating variational graph autoencoder and local-global graph network[J]. Journal of Computer Applications, 2025,45(6):1849-1857.
[99]
DEBIEE, FERNANDEZ ROJASR, FIDOCKJ, et al. Multimodal fusion for objective assessment of cognitive workload: A review[J]. IEEE Transactions on Cybernetics, 2021, 51(3): 1542-1555. DOI:10.1109/TCYB.2019.2939399 .
[100]
XUK M, KOORNP, DE KONINGB, et al. A growth mindset lowers perceived cognitive load and improves learning: Integrating motivation to cognitive load[J]. Journal of Educational Psychology, 2021, 113(6): 1177-1191. DOI:10.1037/edu0000631 .