IEEE Computational Intelligence Magazine - November 2022 - 25

V. Conclusion
Meta-learning is an effective low-resource machine learning
strategy. This paper has proposed a meta-learning approach,
MDMAML, to address challenging cross-subject, few-shot
and source-free (privacy protection) classification tasks in
EEG-based BCIs. It integrates learning-to-learn across multiple
source domains, layer freezing, and negative transfer mitigation.
Experiments on four datasets from two different BCI
paradigms (MI and ERP) demonstrated that MDMAML outperformed
several classical and state-of-the-art approaches in
both online and offline applications.
To our knowledge, this is the first work that addresses
individual differences, data scarcity, and data privacy protection
simultaneously in EEG-based BCIs.
FIGURE 6 ECE and MCE in cross-subject classification on MI1.
(a) MDMAML using EEGNet, (b) MAML using EEGNet, (c) MDMAML
using Shallow ConvNet, and (d) MAML using Shallow ConvNet.
Acknowledgments
This work was supported in part by the Hubei Province
Funds for Distinguished Young Scholars under Grant
2020CFA050, in part by CCF-AFSG Research Fund under
Grant CCF-AFSG RF20210007, in part by the Technology
Innovation Project of Hubei Province of China under Grant
2019AEA171, and in part by the Beijing Fenghuo Wanjia
Technology Co. Ltd.
References
FIGURE 7 Model Inversion Attack onMI1. The victimmodel was EEGNet
trained byMDMAML. Different colors represent samples from different
subjects. Red dots represent 64 reconstructed samples froma particular
batch.
t-SNE [30] visualization of the reconstruction results
on MI1are showninFig.7,where thered dots represent
64 reconstructed samples from a particular batch, and each
remaining color represents data from a different source
subject (Subjects s0, s2, s6, s7 and s8 were used because of
their better performance). The red dots do not resemble
data from any particular source subject; hence, it seems
impossible to discover the identity of a particular source
subject.
[1] A. M. Turing, " Computing machinery and intelligence, " Mind, vol. LIX,
vol. 59, no. 236, pp. 433-460, 1950.
[2] B. Graimann, B. Allison, and G. Pfurtscheller, " Brain-computer interfaces: A
gentle introduction, " Brain- Comput. Interfaces, pp. 1-27, 2009.
[3] D. Wu, Y. Xu, and B.-L. Lu, " Transfer learning for EEG-based brain-computer
interfaces: A review of progress made since 2016, " IEEE Trans. Cogn. Devlop. Syst.,
vol. 14, no. 1, pp. 4-19, Mar. 2022.
[4] B. Z. Allison, E. W. Wolpaw, and J. R. Wolpaw, " Brain-computer interface systems:
Progress and prospects, " Expert Rev. Med. Devices, vol. 4, no. 4, pp. 463-474,
2007.
[5] F. R. Willett, D. T. Avansino, L. R. Hochberg, J. M. Henderson, and K. V. Shenoy,
" High-performance brain-to-text communication via handwriting, " Nature,
vol. 593, no. 7858, pp. 249-254, 2021.
[6] U. Chaudhary et al., " Spelling interface using intracortical signals in a completely
locked-in patient enabled via auditory neurofeedback training, " Nature Commun.,
vol. 13, no. 1, pp. 1-9, 2022.
[7] A. Zimmermann-Schlatter, C. Schuster, M. A. Puhan, E. Siekierka, and J.
Steurer, " Efficacy of motor imagery in post-stroke rehabilitation: A systematic
review, " J. Neuroengineering Rehabil., vol. 5, no. 1, pp. 1-10, 2008.
[8] X. Tian et al., " Deep multi-view feature learning for EEG-based epileptic seizure
detection, " IEEE Trans. Neural Syst. Rehabil. Eng., vol. 27, no. 10, pp. 1962-1972,
Oct. 2019.
[9] R. Peng, J. Jiang, G. Kuang, H. Du, D. Wu, and J. Shao, " EEG-based automatic
epilepsy detection: Review and outlook, " (in Chinese), Acta Automatica Sinica,
vol. 48, no. 2, pp. 335-350, 2022.
[10] M. M. Shanechi, " Brain-machine interfaces from motor to mood, " Nature Neurosci.,
vol. 22, no. 10, pp. 1554-1564, 2019.
[11] D. Wu, V. J. Lawhern, S. Gordon, B. J. Lance, and C.-T. Lin, " Driver drowsiness
estimation from EEG signals using online weighted adaptation regularization for
regression (OwARR), " IEEE Trans. Fuzzy Syst., vol. 25, no. 6, pp. 1522-1535,
Dec. 2017.
[12] V. Mihajlovic, B. Grundlehner, R. Vullers, and J. Penders, " Wearable, wireless
EEG solutions in daily life applications: What are we missing?, " J. Biomed. Health
Inform., vol. 19, no. 1, pp. 6-21, 2014.
[13] J. C. Henry, " Electroencephalography: Basic principles, clinical applications,
and related fields, " Neurology, vol. 67, no. 11, pp. 2092-2092, 2006.
[14] G. Pfurtscheller and C. Neuper, " Motor imagery and direct brain-computer
communication, " Proc. IEEE, vol. 89, no. 7, pp. 1123-1134, Jul. 2001.
[15] T. W. Picton et al., " Guidelines for using human event-related potentials to
study cognition: Recording standards and publication criteria, " Psychophysiol.,
vol. 37, pp. 127-152, 2000.
[16] A. M. Norcia, L. G. Appelbaum, J. M. Ales, B. R. Cottereau, and B. Rossion,
" The steady-state visual evoked potential in vision research: A review, " J. Vis.,
vol. 15, no. 6, p. 4, 2015.
NOVEMBER 2022 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE 25
IEEE Computational Intelligence Magazine - November 2022

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - November 2022
