摘要
脑机接口(Brain Computer Interface, BCI)能够在人脑与外部设备之间建立直接的信息交互通路,而无需依赖外周神经与肌肉。经过50余年的发展,BCI研究已从通讯接口设计和自然交互实现阶段,逐渐向脑机智能融合阶段过渡。然而,实现BCI的长时程稳定交互,成为制约其进一步发展和实际应用的关键瓶颈。目前,提升BCI长期稳定性的技术途径主要包括两个方面:一是通过大脑刺激与神经反馈训练等手段增强用户使用BCI的能力,即脑学习;二是利用自适应技术提高机器端对脑信号的解码与适应能力,即机学习。现有多数研究往往单独关注脑学习或机学习,而忽略了二者结合对性能提升的巨大潜力。协同演进BCI则通过有效融合脑学习和机学习过程,使大脑与机器在交互过程中实现相互适应和同步增强,从而克服大脑状态波动和外部环境干扰,支撑BCI的长时程稳定运行。本文系统梳理了神经调制、神经反馈及自适应BCI技术的发展,明确了协同演进BCI的定义与边界,并进一步分析其面临的关键问题与未来发展方向,以期推动协同演进BCI的深入研究与广泛应用。
Towards Brain-Machine Intelligence: Co-Evolution Brain-Computer Interface
- Haihe Laboratory of Brain-Computer Interaction and Human-Machine Integration, Tianjin 300392, China
Abstract
Brain-computer interfaces (BCIs) enable direct information exchange between the human brain and external devices without relying on peripheral nerves and muscles. Over the past five decades, BCI research has progressed from the design of communication interfaces and natural interaction to the stage of brain-machine intelligence. However, achieving long-term stable interaction remains a major bottleneck limiting further development and practical application of BCI systems. Current approaches to improve long-term stability mainly fall into two categories: enhancing the user’s ability to operate the BCI through neural modulation and neurofeedback training, referred to as brain learning; and improving the machine’s decoding and adaptation to brain signals through adaptive algorithms, referred to as machine learning. Most existing studies tend to focus on either brain learning or machine learning in isolation, overlooking the potential of their integration for performance enhancement. Co-evolution BCI aims to synergise brain learning and machine learning, enabling mutual adaptation and synchronised enhancement between the brain and machine during interaction. This facilitates robust performance against brain state fluctuations and environmental disturbances, thereby supporting long-term stable BCI operation. This paper provides a comprehensive review of the development of neural modulation, neurofeedback, and adaptive BCI technologies, defines the concept and scope of co-evolution BCI, and analyses the key challenges and future directions to promote deeper research and broader application of co-evolution BCIs
参考文献
[1] VIDAL J J. Toward direct brain-computer communication[J]. Annual Review of Biophysics and Bioengineering, 1973, 2(1): 157-180.
[2] MCFARLAND D J. EEG-based brain-computer interfaces[J]. Current Opinion in Biomedical Engineering, 2017, 3: 194-200.
[3] XU M, HE F, JUNG T P, et al. Current challenges for the practical application of electroencephalography-based brain-computer interfaces[J]. Engineering, 2021, 7(12): 1710-1712.
[4] XU M, HAN J, WANG Y, et al. Implementing over 100 command codes for a high-speed hybrid brain-computer interface using concurrent P300 and SSVEP features[J]. IEEE Transactions on Biomedical Engineering, 2020, 67(4): 971-981.
[5] HAN J, XU M, XIAO X, et al. A high-speed hybrid brain-computer interface with more than 200 targets[J]. Journal of Neural Engineering, 2023, 20(1): 016025.
[6] JIANG J, YIN E, WANG C, et al. Incorporation of dynamic stopping strategy into the high-speed SSVEP-based BCIs[J]. Journal of Neural Engineering, 2018, 15(4): 046025.
[7] WILLETT F R, AVANSINO D T, HOCHBERG L R, et al. High-performance brain-to-text communication via handwriting[J]. Nature, 2021, 593(7858): 249-254.
[8] XU M, XIAO X, WANG Y, et al. A brain-computer interface based on miniature event-related potentials induced by very small lateral visual stimuli[J]. IEEE Transactions on Biomedical Engineering, 2018, 65(5): 1166-1175.
[9] MEI J, XU M, WANG L, et al. Using SSVEP-BCI to continuously control a quadcopter with 4-DOF motions[C]//2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC). Montreal, QC, Canada: IEEE, 2020: 4745-4748.
[10] MEI J, CHEN W, LI A, et al. A high-DOF BCI control strategy mapping discrete commands to continuous motion for a drone[J]. IEEE Transactions on Automation Science and Engineering, 2025.
[11] METZGER S L, LITTLEJOHN K T, SILVA A B, et al. A high-performance neuroprosthesis for speech decoding and avatar control[J]. Nature, 2023, 620(7976): 1037-1046.
[12] WILLSEY M S, SHAH N P, AVANSINO D T, et al. A high-performance brain-computer interface for finger decoding and quadcopter game control in an individual with paralysis[J]. Nature Medicine, 2025, 31(1): 96-104.
[13] GAO X, WANG Y, CHEN X, et al. Interface, interaction, and intelligence in generalized brain-computer interfaces[J]. Trends in Cognitive Sciences, 2021, 25(5): 432-452.
[14] PERDIKIS S, MILLÁN J D R. Brain-machine interfaces: a tale of two learners[J]. IEEE Systems, Man, and Cybernetics Magazine, 2020, 6(3): 12-19.
[15] DADARLAT M C, CANFIELD R A, ORSBORN A L. Neural plasticity in sensorimotor brain-machine interfaces[J]. Annual Review of Biomedical Engineering, 2023, 25(1): 51-76.
[16] SORRELL E, RULE M E, O’LEARY T. Brain-machine interfaces: closed-loop control in an adaptive system[J]. Annual Review of Control, Robotics, and Autonomous Systems, 2021, 4(1): 167-189.
[17] LIU Z, MEI J, TANG J, et al. A memristor-based adaptive neuromorphic decoder for brain-computer interfaces[J]. Nature Electronics, 2025, 8(4): 362-372.
[18] BIASIUCCI A, LEEB R, ITURRATE I, et al. Brain-actuated functional electrical stimulation elicits lasting arm motor recovery after stroke[J]. Nature Communications, 2018, 9(1): 2421.
[19] KRUEGER J, KRAUTH R, REICHERT C, et al. Hebbian plasticity induced by temporally coincident BCI enhances post-stroke motor recovery[J]. Scientific Reports, 2024, 14(1): 18700.
[20] ZHANG Y, GAO Y, ZHOU J, et al. Advances in brain-computer interface controlled functional electrical stimulation for upper limb recovery after stroke[J]. Brain Research Bulletin, 2025, 226: 111354.
[21] IWANE F, ITURRATE I, CHAVARRIGA R, et al. Invariability of EEG error-related potentials during continuous feedback protocols elicited by erroneous actions at predicted or unpredicted states[J]. Journal of Neural Engineering, 2021, 18(4): 046044.
[22] SPRIGGS M J, SUMNER R L, MCMILLAN R L, et al. Indexing sensory plasticity: evidence for distinct predictive coding and Hebbian learning mechanisms in the cerebral cortex[J]. NeuroImage, 2018, 176: 290-300.
[23] BENNETT D, NIV Y, LANGDON A J. Value-free reinforcement learning: policy optimization as a minimal model of operant behavior[J]. Current Opinion in Behavioral Sciences, 2021, 41: 114-121.
[24] DEROSIERE G, SHOKUR S, VASSILIADIS P. Reward signals in the motor cortex: from biology to neurotechnology[J]. Nature Communications, 2025, 16(1): 1307.
[25] VELLISTE M, PEREL S, SPALDING M C, et al. Cortical control of a prosthetic arm for self-feeding[J]. Nature, 2008, 453(7198): 1098-1101.
[26] BOUTON C E, SHAIKHOUNI A, ANNETTA N V, et al. Restoring cortical control of functional movement in a human with quadriplegia[J]. Nature, 2016, 533(7602): 247-250.
[27] EDELMAN B J, MENG J, SUMA D, et al. Noninvasive neuroimaging enhances continuous neural tracking for robotic device control[J]. Science Robotics, 2019, 4(31): eaaw6844.
[28] WILLSEY M S, SHAH N P, AVANSINO D T, et al. A high-performance brain-computer interface for finger decoding and quadcopter game control in an individual with paralysis[J]. Nature Medicine, 2025, 31(1): 96-104.
[29] HERRON J, KREMEN V, SIMERAL J D, et al. The convergence of neuromodulation and brain-computer interfaces[J]. Nature Reviews Bioengineering, 2024, 2(8): 628-630.
[30] KRAMES E S, PECKHAM P H, REZAI A R, et al. Neuromodulation: comprehensive textbook of principles, technologies, and therapies[M]. 2nd ed. London: Academic Press, 2018.
[31] LIU A, VÖRÖSLAROS M, KRONBERG G, et al. Immediate neurophysiological effects of transcranial electrical stimulation[J]. Nature Communications, 2018, 9(1): 5092.
[32] ZHONG G, YANG Z, JIANG T. Precise modulation strategies for transcranial magnetic stimulation: advances and future directions[J]. Neuroscience Bulletin, 2021, 37(12): 1718-1734.
[33] KUBANEK J. Neuromodulation with transcranial focused ultrasound[J]. Neurosurgical Focus, 2018, 44(2): E14.
[34] LEGON W, SATO T F, OPITZ A, et al. Transcranial focused ultrasound modulates the activity of primary somatosensory cortex in humans[J]. Nature Neuroscience, 2014, 17(2): 322-329.
[35] LOZANO A M, LIPSMAN N, BERGMAN H, et al. Deep brain stimulation: current challenges and future directions[J]. Nature Reviews Neurology, 2019, 15(3): 148-160.
[36] KRAUSS J K, LIPSMAN N, AZIZ T, et al. Technology of deep brain stimulation: current status and future directions[J]. Nature Reviews Neurology, 2021, 17(2): 75-87.
[37] ZAIMI B, TURNBULL M, HAZRA A, et al. Closed-loop optogenetic control of the dynamics of neural activity in non-human primates[J]. Nature Biomedical Engineering, 2022, 7(4): 559-575.
[38] BANSAL A, SHIKHA S, ZHANG Y. Towards translational optogenetics[J]. Nature Biomedical Engineering, 2022, 7(4): 349-369.
[39] ZHOU A, SANTACRUZ S R, JOHNSON B C, et al. A wireless and artefact-free 128-channel neuromodulation device for closed-loop stimulation and recording in non-human primates[J]. Nature Biomedical Engineering, 2018, 3(1): 15-26.
[40] FLESHER S N, COLLINGER J L, FOLDES S T, et al. Intracortical microstimulation of human somatosensory cortex[J]. Science Translational Medicine, 2016, 8(361): 361ra141.
[41] LI Z, ZHANG R, LI W, et al. Enhancement of hybrid BCI system performance based on motor imagery and SSVEP by transcranial alternating current stimulation[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2024, 32: 3222-3230.
[42] ONAGAWA R, MURAOKA Y, HAGURA N, et al. An investigation of the effectiveness of neurofeedback training on motor performance in healthy adults: a systematic review and meta-analysis[J]. NeuroImage, 2023, 270: 120000.
[43] SITARAM R, ROS T, STOECKEL L, et al. Closed-loop brain training: the science of neurofeedback[J]. Nature Reviews Neuroscience, 2017, 18(2): 86-100.
[44] BEATTY J, GREENBERG A, DEIBLER W P, et al. Operant control of occipital theta rhythm affects performance in a radar monitoring task[J]. Science, 1974, 183(4127): 871-873.
[45] ROS T, THÉBERGE J, FREWEN P A, et al. Mind over chatter: plastic up-regulation of the fMRI salience network directly after EEG neurofeedback[J]. NeuroImage, 2013, 65: 324-335.
[46] HUANG W, WU W, LUCAS M V, et al. Neurofeedback training with an electroencephalogram-based brain-computer interface enhances emotion regulation[J]. IEEE Transactions on Affective Computing, 2023, 14(2): 998-1011.
[47] GRIGOREV N A, SAVOSENKOV A O, LUKOYANOV M V, et al. A BCI-based vibrotactile neurofeedback training improves motor cortical excitability during motor imagery[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2021, 29: 1583-1592.
[48] NOBLE S C, WOODS E, WARD T, et al. Accelerating P300-based neurofeedback training for attention enhancement using iterative learning control: a randomised controlled trial[J]. Journal of Neural Engineering, 2024, 21(2): 026006.
[49] ZHONG Y, WANG Y, FARINA D, et al. A closed-loop tactile stimulation training protocol for motor imagery-based BCI: boosting BCI performance for BCI-deficiency users[J]. IEEE Transactions on Biomedical Engineering, 2025.
[50] DADARLAT M C, O’DOHERTY J E, SABES P N. A learning-based approach to artificial sensory feedback leads to optimal integration[J]. Nature Neuroscience, 2015, 18(1): 138-144.
[51] WAN C, PEI M, SHI K, et al. Toward a brain-neuromorphics interface[J]. Advanced Materials, 2024, 36(37): 2311288.
[52] LIU Z, TANG J, GAO B, et al. Neural signal analysis with memristor arrays towards high-efficiency brain-machine interfaces[J]. Nature Electronics, 2025.
[53] YUAN R, TIW P J, CAI L, et al. A neuromorphic physiological signal processing system based on VO2 memristor for next-generation human-machine interface[J]. Nature Communications, 2023, 14(1): 3695.
[54] PEI M, ZHU Y, LIU S, et al. Power-efficient multisensory reservoir computing based on Zr-doped HfO2 memcapacitive synapse arrays[J]. Advanced Materials, 2023, 35(41): 2305609.
[55] RONCA V, FLUMERI G D, GIORGI A, et al. o-CLEAN: a novel multi-stage algorithm for the ocular artifacts’ correction from EEG data in out-of-the-lab applications[J]. Journal of Neural Engineering, 2024, 21(5): 056023.
[56] GUARNIERI R, MARINO M, BARBAN F, et al. Online EEG artifact removal for BCI applications by adaptive spatial filtering[J]. Journal of Neural Engineering, 2018, 15(5): 056009.
[57] HE C, LIU J, ZHU Y, et al. Data augmentation for deep neural networks model in EEG classification task: a review[J]. Frontiers in Human Neuroscience, 2021, 15: 765525.
[58] WONG C M, WANG Z, NAKANISHI M, et al. Online adaptation boosts SSVEP-based BCI performance[J]. IEEE Transactions on Biomedical Engineering, 2022, 69(6): 2018-2028.
[59] SILVERSMITH D B, ABIRI R, HARDY N F, et al. Plug-and-play control of a brain-computer interface through neural map stabilization[J]. Nature Biotechnology, 2021, 39(3): 326-335.
[60] CHIANG K J, EMMANOUILIDOU D, GAMPER H, et al. A closed-loop adaptive brain-computer interface framework: improving the classifier with the use of error-related potentials[C]//2021 10th International IEEE/EMBS Conference on Neural Engineering (NER). Italy: IEEE, 2021: 487-490.
[61] POHLMEYER E A, MAHMOUDI B, GENG S, et al. Using reinforcement learning to provide stable brain-machine interface control despite neural input reorganization[J]. PLoS ONE, 2014, 9(1): e87253.
[62] MARSH B T, TARIGOPPULA V S A, CHEN C, et al. Toward an autonomous brain machine interface: integrating sensorimotor reward modulation and reinforcement learning[J]. The Journal of Neuroscience, 2015, 35(19): 7374-7387.
[63] SHEN X, ZHANG X, HUANG Y, et al. Intermediate sensory feedback assisted multi-step neural decoding for reinforcement learning based brain-machine interfaces[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022, 30: 2834-2844.
[64] LUO T J, FAN Y C, LV J T, et al. Deep reinforcement learning from error-related potentials via an EEG-based brain-computer interface[C]//2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). Madrid, Spain: IEEE, 2018: 697-701.
[65] GEHRKE L, LOPES P, KLUG M, et al. Neural sources of prediction errors detect unrealistic VR interactions[J]. Journal of Neural Engineering, 2022, 19(3): 036002.
[66] KATO K, SAWADA M, NISHIMURA Y. Bypassing stroke-damaged neural pathways via a neural interface induces targeted cortical adaptation[J]. Nature Communications, 2019, 10(1): 4699.
[67] ZHANG H, CHAVARRIGA R, KHALILIARDALI Z, et al. EEG-based decoding of error-related brain activity in a real-world driving task[J]. Journal of Neural Engineering, 2015, 12(6): 066028.
[68] SALAZAR-GOMEZ A F, DELPRETO J, GIL S, et al. Correcting robot mistakes in real time using EEG signals[C]//2017 IEEE International Conference on Robotics and Automation (ICRA). Singapore: IEEE, 2017: 6570-6577.
[69] EHRLICH S K, CHENG G. A feasibility study for validating robot actions using EEG-based error-related potentials[J]. International Journal of Social Robotics, 2019, 11(2): 271-283.
[70] CRUZ A, PIRES G, NUNES U J. Double ErrP detection for automatic error correction in an ERP-based BCI speller[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2018, 26(1): 26-36.
[71] SORIANO-SEGURA P, ORTIZ M, IÁÑEZ E, et al. Design of a brain-machine interface for reducing false activations of a lower-limb exoskeleton based on error related potential[J]. Computer Methods and Programs in Biomedicine, 2024, 255: 108332.
[72] DEMCHENKO I, SHAVIT T, BENYAMINI M, et al. Self-correcting brain computer interface based on classification of multiple error-related potentials[J]. Journal of Neural Engineering, 2025.
[73] CHIANG K J, EMMANOUILIDOU D, GAMPER H, et al. A closed-loop adaptive brain-computer interface framework: improving the classifier with the use of error-related potentials[C]//2021 10th International IEEE/EMBS Conference on Neural Engineering (NER). Italy: IEEE, 2021: 487-490.
[74] SHIN J H, KWON J, KIM J U, et al. Wearable EEG electronics for a Brain-AI closed-loop system to enhance autonomous machine decision-making[J]. npj Flexible Electronics, 2022, 6(1): 32.
[75] BATZIANOULIS I, IWANE F, WEI S, et al. Customizing skills for assistive robotic manipulators, an inverse reinforcement learning approach with error-related potentials[J]. Communications Biology, 2021, 4(1): 1406.
[76] HAUSER T U, IANNACCONE R, STÄMPFLI P, et al. The feedback-related negativity (FRN) revisited: new insights into the localization, meaning and network organization[J]. NeuroImage, 2014, 84: 159-168.
[77] SUGIMOTO F, KIMURA M, TAKEDA Y. Attenuation of auditory Tb and N200 in response to self-modulated tones during continuous actions reflects the prediction of action consequences[J]. International Journal of Psychophysiology, 2025, 209: 112522.
[78] YANG C, HAN X, WANG Y, et al. A dynamic window recognition algorithm for SSVEP-based brain-computer interfaces using a spatio-temporal equalizer[J]. International Journal of Neural Systems, 2018, 28(10): 1850028.
[79] CHEN Y, YANG C, CHEN X, et al. A novel training-free recognition method for SSVEP-based BCIs using dynamic window strategy[J]. Journal of Neural Engineering, 2021, 18(3): 036007.
[80] JIN J, HE X, ALLISON B Z, et al. Leveraging spatiotemporal estimation for online adaptive steady-state visual evoked potential recognition[J]. IEEE Transactions on Cognitive and Developmental Systems, 2024, 16(6): 1943-1954.
[81] FORENZO D. Continuous tracking using deep learning-based decoding for noninvasive brain-computer interface[J]. PNAS Nexus, 2024, 3(4): pgae145.
[82] SCHWEMMER M A, SKOMROCK N D, SEDERBERG P B, et al. Meeting brain-computer interface user performance expectations using a deep neural network decoding framework[J]. Nature Medicine, 2018, 24(11): 1669-1676.
[83] ZHANG S, YOSHIDA W, MANO H, et al. Pain control by co-adaptive learning in a brain-machine interface[J]. Current Biology, 2020, 30(20): 3935-3944.e7.
[84] WANG J, BI L, FEI W. Multitask-oriented brain-controlled intelligent vehicle based on human-machine intelligence integration[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2023, 53(4): 2510-2521.
[85] ARICÒ P, BORGHINI G, DI FLUMERI G, et al. Adaptive automation triggered by EEG-based mental workload index: a passive brain-computer interface application in realistic air traffic control environment[J]. Frontiers in Human Neuroscience, 2016, 10: 539.
[86] LI M, LI W, NIU L, et al. An event-related potential-based adaptive model for telepresence control of humanoid robot motion in an environment cluttered with obstacles[J]. IEEE Transactions on Industrial Electronics, 2017, 64(2): 1696-1705.
[87] BENAROCH C, SADATNEJAD K, ROC A, et al. Long-term BCI training of a tetraplegic user: adaptive Riemannian classifiers and user training[J]. Frontiers in Human Neuroscience, 2021, 15: 635193.
[88] CHOWDHURY A, RAZA H, MEENA Y K, et al. Online covariate shift detection-based adaptive brain-computer interface to trigger hand exoskeleton feedback for neuro-rehabilitation[J]. IEEE Transactions on Cognitive and Developmental Systems, 2018, 10(4): 1070-1080.
[89] SHANECHI M M, ORSBORN A L, CARMENA J M. Robust brain-machine interface design using optimal feedback control modeling and adaptive point process filtering[J]. PLOS Computational Biology, 2016, 12(4): e1004730.
[90] RIZZOGLIO F, CASADIO M, DE SANTIS D, et al. Building an adaptive interface via unsupervised tracking of latent manifolds[J]. Neural Networks, 2021, 137: 174-187.
[91] KILICARSLAN A, GROSSMAN R G, CONTRERAS-VIDAL J L. A robust adaptive denoising framework for real-time artifact removal in scalp EEG measurements[J]. Journal of Neural Engineering, 2016, 13(2): 026013.
[92] ROY A M. Adaptive transfer learning-based multiscale feature fused deep convolutional neural network for EEG MI multiclassification in brain-computer interface[J]. Engineering Applications of Artificial Intelligence, 2022, 116: 105347.
[93] TSAI B Y, DIDDI S V S, KO L W, et al. Development of an adaptive artifact subspace reconstruction based on Hebbian/anti-Hebbian learning networks for enhancing BCI performance[J]. IEEE Transactions on Neural Networks and Learning Systems, 2024, 35(1): 348-361.
[94] NAROS G, NAROS I, GRIMM F, et al. Reinforcement learning of self-regulated sensorimotor β-oscillations improves motor performance[J]. NeuroImage, 2016, 134: 142-152.
[95] LUO J, CUI W, XU S, et al. A dual-branch spatio-temporal-spectral transformer feature fusion network for EEG-based visual recognition[J]. IEEE Transactions on Industrial Informatics, 2024, 20(2): 1721-1731.
[96] HUANG Y, HE F, XU M, et al. Operate P300 speller when performing other task[J]. Journal of Neural Engineering, 2020, 17(5): 056012.
[97] WANG J, CUI Y, ZHANG H, et al. An asynchronous training-free SSVEP-BCI detection algorithm for non-equal prior probability scenarios[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2024, 32: 4120-4130.
[98] KLEIH S C, KÜBLER A. Psychological factors influencing brain-computer interface (BCI) performance[C]//2015 IEEE International Conference on Systems, Man, and Cybernetics. Kowloon, China: IEEE, 2015: 3192-3196.
[99] DIAS S B, JELINEK H F, HADJILEONTIADIS L J. Multisensed emotions as adaptation controllers in human-to-serious neurogames communication[J]. IEEE Communications Magazine, 2023, 61(10): 38-44.
[100] CHEN Y, PENG Y, TANG J, et al. EEG-based affective brain-computer interfaces: recent advancements and future challenges[J]. Journal of Neural Engineering, 2025.
[101] HAN Y, ZIEBELL P, RICCIO A, et al. Two sides of the same coin: adaptation of BCIs to internal states with user-centered design and electrophysiological features[J]. Brain-Computer Interfaces, 2022, 9(2): 102-114.
[102] HUSSAIN S A H, RAZA I, HUSSAIN S A, et al. A mental state aware brain computer interface for adaptive control of electric powered wheelchair[J]. Scientific Reports, 2025, 15(1): 9880.
