AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals

Syed Mohamed Syed Abubakar Shaihan

doi:doi:10.11648/j.ejcbs.20261203.12

Research Article |

| Peer-Reviewed

AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals

Syed Mohamed Syed Abubakar Shaihan^*

Published in European Journal of Clinical and Biomedical Sciences (Volume 12, Issue 3)

Received: 5 April 2026 Accepted: 20 April 2026 Published: 27 August 2026

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

In this paper, we discuss the recognition of electroencephalographic (EEG) signals, which is crucial in order to improve the performance of non-invasive brain-computer interfaces (BCIs). Although deep learning (DL) has achieved considerable advancements in the decoding of EEG signals, it frequently encounters difficulties related to noisy data and non-stationarity challenges. We discuss geometric learning which offers a more robust way to handle EEG signals by leveraging the mathematical structure of the data. We extended the existing Manifold Attention Network (MAtt), a novel deep learning model that applies a manifold attention mechanism to better capture the spatiotemporal patterns of EEG signals. Instead of using traditional Euclidean space, we mapped the data onto a Riemannian symmetric positive definite (SPD) manifold, which allows for more effective feature extraction. One major issue with existing SPD-based deep learning approaches is that they rely on fixed Riemannian metrics, which can be suboptimal. To solve this, we integrate Adaptive Log-Euclidean Metrics (ALEMs)---a learnable metric framework into the MAtt network that adapts to the specific structure of EEG data, improving model performance with minimal extra computation. AMAtt achieves 63.19% accuracy on BCIC-IV-2a and 46.00% on MAMEM-SSVEP-II, outperforming the MAtt baseline by 5.55% and 4.60% respectively. This adaptive geometric approach opens new possibilities for robust, subject-independent EEG decoding, paving the way towards practical BCI systems.

Published in	European Journal of Clinical and Biomedical Sciences (Volume 12, Issue 3)
DOI	10.11648/j.ejcbs.20261203.12
Page(s)	54-61
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2026. Published by Science Publishing Group

Keywords

EEG, Manifold Attention Network, Adaptive Log-Euclidean, SPD, BCI

References

[1]	Pan, Y.-T.; Chou, J.-L.; Wei, C.-S. MAtt: A Manifold Attention Network for EEG Decoding. Advances in Neural Information Processing Systems 2022, 35, pp. 31116-31129. https://doi.org/10.48550/arXiv.2210.01986
[2]	Iturrate, I.; Antelis, J.; Minguez, J. Synchronous EEG brain-actuated wheelchair with automated navigation. In Proceedings of the 2009 IEEE International Conference on Robotics and Automation, 2009. https://doi.org/10.1109/ROBOT.2009.5152580
[3]	Subha, D. P.; Joseph, P. K.; Acharya U, R.; Lim, C. M. EEG signal analysis: a survey. Journal of Medical Systems 2010, 34(2). https://doi.org/10.1007/s10916-008-9231-z
[4]	Al-Qaysi, Z. T.; Zaidan, B. B.; Zaidan, A. A.; Suzani, M. S. A review of disability EEG based wheelchair control system: Coherent taxonomy, open challenges and recommendations. Computer Methods and Programs in Biomedicine 2018, 164. https://doi.org/10.1016/j.cmpb.2018.06.012
[5]	Chen, X.; Wang, Y.; Nakanishi, M.; Gao, X.; Jung, T.-P.; Gao, S. High-speed spelling with a noninvasive brain-computer interface. Proceedings of the National Academy of Sciences 2015, 112(44). https://doi.org/10.1073/pnas.1508080112
[6]	Barachant, A.; Bonnet, S.; Congedo, M.; Jutten, C. Multiclass brain-computer interface classification by Riemannian geometry. IEEE Transactions on Biomedical Engineering 2012, 59(4), pp. 920-928. https://doi.org/10.1109/TBME.2011.2172210
[7]	Monti, F.; Boscaini, D.; Masci, J.; Rodola, E.; Svoboda, J.; Bronstein, M. M. Geometric deep learning on graphs and manifolds using mixture model CNNs. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017. https://doi.org/10.48550/arXiv.1611.08402
[8]	Chen, Z.; Song, Y.; Xu, T.; Huang, Z.; Wu, X.-J.; Sebe, N. Adaptive Log-Euclidean Metrics for SPD Matrix Learning. IEEE Transactions on Image Processing 2024, 33, pp. 5194-5205. https://doi.org/10.1109/TIP.2024.3451930
[9]	Arsigny, V.; Fillard, P.; Pennec, X.; Ayache, N. Log-Euclidean metrics for fast and simple calculus on diffusion tensors. Magnetic Resonance in Medicine 2006, 56(2), pp. 411-421. https://doi.org/10.1002/mrm.20965
[10]	Barbaresco, F. Innovative tools for radar signal processing based on Cartan's geometry of SPD matrices & information geometry. In Proceedings of the 2008 IEEE Radar Conference, 2008. https://doi.org/10.1109/RADAR.2008.4720937
[11]	Arsigny, V.; Fillard, P.; Pennec, X.; Ayache, N. Geometric means in a novel vector space structure on symmetric positive-definite matrices. SIAM Journal on Matrix Analysis and Applications 2007, 29(1), pp. 328-347. https://doi.org/10.1137/050637996
[12]	Nguyen, X. S. A Gyrovector space approach for symmetric positive semi-definite matrix learning. In Proceedings of the European Conference on Computer Vision, 2022; pp. 52-68. https://doi.org/10.1007/978-3-031-19812-0_4
[13]	Brunner, C.; Leeb, R.; M"{u}ller-Putz, G.; Schl"{o}gl, A.; Pfurtscheller, G. BCI Competition 2008-Graz data set A. Institute for Knowledge Discovery (Laboratory of Brain-Computer Interfaces), Graz University of Technology 2008, 16.
[14]	Dimitriadis, S. I. SCZ: A Riemannian schizophrenia diagnosis framework based on the multiplexity of EEG-based dynamic functional connectivity patterns. Computers in Biology and Medicine 2024, 180. https://doi.org/10.1016/j.compbiomed.2024.108862
[15]	Lawhern, V. J.; Solon, A. J.; Waytowich, N. R.; Gordon, S. M.; Hung, C. P.; Lance, B. J. EEGNet: A compact convolutional neural network for EEG-based brain-computer interfaces. Journal of Neural Engineering 2018, 15(5), 056013. https://doi.org/10.1088/1741-2552/aace8c
[16]	Schirrmeister, R. T.; Springenberg, J. T.; Fiederer, L. D. J.; Glasstetter, M.; Eggensperger, K.; Tangermann, M.; Hutter, F.; Burgard, W.; Ball, T. Deep learning with convolutional neural networks for EEG decoding and visualization. Human Brain Mapping 2017, 38(11), pp. 5391-5420. https://doi.org/10.1002/hbm.23730

Cite This Article

Plain Text BibTeX RIS

APA Style

Shaihan, S. M. S. A. (2026). AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. European Journal of Clinical and Biomedical Sciences, 12(3), 54-61. https://doi.org/10.11648/j.ejcbs.20261203.12

Copy | Download

ACS Style

Shaihan, S. M. S. A. AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. Eur. J. Clin. Biomed. Sci. 2026, 12(3), 54-61. doi: 10.11648/j.ejcbs.20261203.12

Copy | Download

AMA Style

Shaihan SMSA. AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. Eur J Clin Biomed Sci. 2026;12(3):54-61. doi: 10.11648/j.ejcbs.20261203.12

Copy | Download

@article{10.11648/j.ejcbs.20261203.12,
  author = {Syed Mohamed Syed Abubakar Shaihan},
  title = {AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals},
  journal = {European Journal of Clinical and Biomedical Sciences},
  volume = {12},
  number = {3},
  pages = {54-61},
  doi = {10.11648/j.ejcbs.20261203.12},
  url = {https://doi.org/10.11648/j.ejcbs.20261203.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ejcbs.20261203.12},
  abstract = {In this paper, we discuss the recognition of electroencephalographic (EEG) signals, which is crucial in order to improve the performance of non-invasive brain-computer interfaces (BCIs). Although deep learning (DL) has achieved considerable advancements in the decoding of EEG signals, it frequently encounters difficulties related to noisy data and non-stationarity challenges. We discuss geometric learning which offers a more robust way to handle EEG signals by leveraging the mathematical structure of the data. We extended the existing Manifold Attention Network (MAtt), a novel deep learning model that applies a manifold attention mechanism to better capture the spatiotemporal patterns of EEG signals. Instead of using traditional Euclidean space, we mapped the data onto a Riemannian symmetric positive definite (SPD) manifold, which allows for more effective feature extraction. One major issue with existing SPD-based deep learning approaches is that they rely on fixed Riemannian metrics, which can be suboptimal. To solve this, we integrate Adaptive Log-Euclidean Metrics (ALEMs)---a learnable metric framework into the MAtt network that adapts to the specific structure of EEG data, improving model performance with minimal extra computation. AMAtt achieves 63.19% accuracy on BCIC-IV-2a and 46.00% on MAMEM-SSVEP-II, outperforming the MAtt baseline by 5.55% and 4.60% respectively. This adaptive geometric approach opens new possibilities for robust, subject-independent EEG decoding, paving the way towards practical BCI systems.},
 year = {2026}
}

Copy | Download

TY  - JOUR
T1  - AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals
AU  - Syed Mohamed Syed Abubakar Shaihan
Y1  - 2026/08/27
PY  - 2026
N1  - https://doi.org/10.11648/j.ejcbs.20261203.12
DO  - 10.11648/j.ejcbs.20261203.12
T2  - European Journal of Clinical and Biomedical Sciences
JF  - European Journal of Clinical and Biomedical Sciences
JO  - European Journal of Clinical and Biomedical Sciences
SP  - 54
EP  - 61
PB  - Science Publishing Group
SN  - 2575-5005
UR  - https://doi.org/10.11648/j.ejcbs.20261203.12
AB  - In this paper, we discuss the recognition of electroencephalographic (EEG) signals, which is crucial in order to improve the performance of non-invasive brain-computer interfaces (BCIs). Although deep learning (DL) has achieved considerable advancements in the decoding of EEG signals, it frequently encounters difficulties related to noisy data and non-stationarity challenges. We discuss geometric learning which offers a more robust way to handle EEG signals by leveraging the mathematical structure of the data. We extended the existing Manifold Attention Network (MAtt), a novel deep learning model that applies a manifold attention mechanism to better capture the spatiotemporal patterns of EEG signals. Instead of using traditional Euclidean space, we mapped the data onto a Riemannian symmetric positive definite (SPD) manifold, which allows for more effective feature extraction. One major issue with existing SPD-based deep learning approaches is that they rely on fixed Riemannian metrics, which can be suboptimal. To solve this, we integrate Adaptive Log-Euclidean Metrics (ALEMs)---a learnable metric framework into the MAtt network that adapts to the specific structure of EEG data, improving model performance with minimal extra computation. AMAtt achieves 63.19% accuracy on BCIC-IV-2a and 46.00% on MAMEM-SSVEP-II, outperforming the MAtt baseline by 5.55% and 4.60% respectively. This adaptive geometric approach opens new possibilities for robust, subject-independent EEG decoding, paving the way towards practical BCI systems.
VL  - 12
IS  - 3
ER  -

Copy | Download

Author Information

Syed Mohamed Syed Abubakar Shaihan

Department of Computer Science, Control and Management, Sapienza University of Rome, Rome, Italy; Department of Mechanical Engineering, Indian Institute of Science, Bangalore, India; Department of Electronics and Communication Engineering, Anjuman Institute of Technology and Management, Bhatkal, India

Contact Email

http://orcid.org/0000-0002-4119-0045

Download PDF

Submit an Article

Sections

Plain Text BibTeX RIS

APA Style

Shaihan, S. M. S. A. (2026). AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. European Journal of Clinical and Biomedical Sciences, 12(3), 54-61. https://doi.org/10.11648/j.ejcbs.20261203.12

Copy | Download

ACS Style

Shaihan, S. M. S. A. AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. Eur. J. Clin. Biomed. Sci. 2026, 12(3), 54-61. doi: 10.11648/j.ejcbs.20261203.12

Copy | Download

AMA Style

Shaihan SMSA. AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals. Eur J Clin Biomed Sci. 2026;12(3):54-61. doi: 10.11648/j.ejcbs.20261203.12

Copy | Download

@article{10.11648/j.ejcbs.20261203.12,
  author = {Syed Mohamed Syed Abubakar Shaihan},
  title = {AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals},
  journal = {European Journal of Clinical and Biomedical Sciences},
  volume = {12},
  number = {3},
  pages = {54-61},
  doi = {10.11648/j.ejcbs.20261203.12},
  url = {https://doi.org/10.11648/j.ejcbs.20261203.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ejcbs.20261203.12},
  abstract = {In this paper, we discuss the recognition of electroencephalographic (EEG) signals, which is crucial in order to improve the performance of non-invasive brain-computer interfaces (BCIs). Although deep learning (DL) has achieved considerable advancements in the decoding of EEG signals, it frequently encounters difficulties related to noisy data and non-stationarity challenges. We discuss geometric learning which offers a more robust way to handle EEG signals by leveraging the mathematical structure of the data. We extended the existing Manifold Attention Network (MAtt), a novel deep learning model that applies a manifold attention mechanism to better capture the spatiotemporal patterns of EEG signals. Instead of using traditional Euclidean space, we mapped the data onto a Riemannian symmetric positive definite (SPD) manifold, which allows for more effective feature extraction. One major issue with existing SPD-based deep learning approaches is that they rely on fixed Riemannian metrics, which can be suboptimal. To solve this, we integrate Adaptive Log-Euclidean Metrics (ALEMs)---a learnable metric framework into the MAtt network that adapts to the specific structure of EEG data, improving model performance with minimal extra computation. AMAtt achieves 63.19% accuracy on BCIC-IV-2a and 46.00% on MAMEM-SSVEP-II, outperforming the MAtt baseline by 5.55% and 4.60% respectively. This adaptive geometric approach opens new possibilities for robust, subject-independent EEG decoding, paving the way towards practical BCI systems.},
 year = {2026}
}

Copy | Download

TY  - JOUR
T1  - AMAtt: Manifold Attention Network with Adaptive Log-Euclidean Metrics for Brain Signals
AU  - Syed Mohamed Syed Abubakar Shaihan
Y1  - 2026/08/27
PY  - 2026
N1  - https://doi.org/10.11648/j.ejcbs.20261203.12
DO  - 10.11648/j.ejcbs.20261203.12
T2  - European Journal of Clinical and Biomedical Sciences
JF  - European Journal of Clinical and Biomedical Sciences
JO  - European Journal of Clinical and Biomedical Sciences
SP  - 54
EP  - 61
PB  - Science Publishing Group
SN  - 2575-5005
UR  - https://doi.org/10.11648/j.ejcbs.20261203.12
AB  - In this paper, we discuss the recognition of electroencephalographic (EEG) signals, which is crucial in order to improve the performance of non-invasive brain-computer interfaces (BCIs). Although deep learning (DL) has achieved considerable advancements in the decoding of EEG signals, it frequently encounters difficulties related to noisy data and non-stationarity challenges. We discuss geometric learning which offers a more robust way to handle EEG signals by leveraging the mathematical structure of the data. We extended the existing Manifold Attention Network (MAtt), a novel deep learning model that applies a manifold attention mechanism to better capture the spatiotemporal patterns of EEG signals. Instead of using traditional Euclidean space, we mapped the data onto a Riemannian symmetric positive definite (SPD) manifold, which allows for more effective feature extraction. One major issue with existing SPD-based deep learning approaches is that they rely on fixed Riemannian metrics, which can be suboptimal. To solve this, we integrate Adaptive Log-Euclidean Metrics (ALEMs)---a learnable metric framework into the MAtt network that adapts to the specific structure of EEG data, improving model performance with minimal extra computation. AMAtt achieves 63.19% accuracy on BCIC-IV-2a and 46.00% on MAMEM-SSVEP-II, outperforming the MAtt baseline by 5.55% and 4.60% respectively. This adaptive geometric approach opens new possibilities for robust, subject-independent EEG decoding, paving the way towards practical BCI systems.
VL  - 12
IS  - 3
ER  -

Copy | Download