Pith. sign in

REVIEW 3 major objections 1 minor 38 references

A network that maps EEG signals to frequency-specific Riemannian manifolds and adds intra-inter slice attention improves cross-subject mental stress detection.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.3

2026-07-03 21:43 UTC pith:WHSW3TM2

load-bearing objection I²RiMA adds per-frequency SPD covariances, data-driven frequency clustering, and intra-inter attention to EEG stress detection, but the abstract gives no experimental details to judge whether the gains are real. the 3 major comments →

arxiv 2607.01279 v1 pith:WHSW3TM2 submitted 2026-07-01 cs.LG

Itextsuperscript{2}RiMA: Spectral Riemannian Representation with Temporal Attention for Mental Stress Detection based on EEG Signals

Cheng He , Kunyu Peng , Shangen Han , Jinming Ma , Jinhong Ding , Likun Xia This is my paper
classification cs.LG
keywords EEGmental stress detectionRiemannian manifoldattention mechanismcross-subject classificationspectral covariancetemporal attention
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tries to show that building spatial covariance matrices at individual frequency points, mapping them into the tangent space, then grouping those frequencies into data-driven clusters and applying attention across time slices yields better detection of stress patterns that vary between people. Conventional approaches either ignore frequency content or break up the temporal flow of EEG recordings, so the new components aim to keep both the geometric structure of brain activity and its rhythmic organization intact. A sympathetic reader would care because reliable cross-subject performance could reduce the need for per-person calibration in applications like workplace monitoring or clinical assessment. If the claim holds, stress detection becomes more accurate and computationally light without sacrificing the ability to handle frequency-specific neural oscillations.

Core claim

I²RiMA constructs spatial covariance matrices independently at each frequency point and maps them to the SPD tangent space, preserving channel-wise geometry together with frequency-specific discriminative cues. It further introduces frequency cluster aggregation to select informative spectral components and reduce redundancy by forming compact, data-driven frequency clusters aligned with EEG rhythms. Finally, an intra-inter slice attention module adaptively integrates local slice-level spectral dynamics and global temporal context across EEG sequences.

What carries the argument

Frequency cluster aggregation combined with the intra-inter slice attention module operating on tangent-space representations of frequency-specific covariance matrices.

Load-bearing premise

The frequency clusters and attention patterns learned on the training subjects capture stress-related neural activity that generalizes to new people rather than dataset-specific noise.

What would settle it

Performance on a held-out fourth EEG dataset collected under different hardware or subject demographics drops below the best baseline while the reported efficiency numbers remain unchanged.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Balanced accuracy reaches up to 82.78 percent on three public datasets.
  • Model size stays at 1.60 million parameters and 31.95 million FLOPs.
  • The method outperforms five existing state-of-the-art approaches under cross-subject evaluation.
  • Frequency-specific geometry is retained without exploding computational cost.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same frequency-cluster plus attention design could be tested on other EEG classification tasks such as drowsiness or emotion recognition.
  • Low parameter count suggests the architecture may support on-device inference for wearable stress monitors.
  • If frequency clusters align with standard EEG bands, the method might reduce the need for manual band selection in future studies.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 1 minor

Summary. The manuscript proposes I²RiMA, an Intra-Inter Riemannian Manifold Attention Network for cross-subject EEG-based mental stress detection. It constructs per-frequency spatial covariance matrices on the SPD manifold, maps them to the tangent space, applies frequency cluster aggregation to form data-driven spectral groups, and uses an intra-inter slice attention module to combine local spectral dynamics with global temporal context. Experiments on three datasets are reported to show consistent outperformance over five state-of-the-art baselines, reaching up to 82.78% balanced accuracy with 1.60M parameters and 31.95M FLOPs.

Significance. If the empirical gains prove robust under proper subject-independent validation, the combination of frequency-aware Riemannian geometry with adaptive temporal attention could advance practical EEG stress detection by better preserving both spatial geometry and oscillatory information while remaining computationally light.

major comments (3)
  1. [Experiments] Experiments section: performance figures (including the 82.78% peak) are stated without error bars, statistical tests, or explicit description of the cross-validation protocol (e.g., leave-one-subject-out or stratified subject splits), which is load-bearing for the central claim of consistent outperformance and generalizability.
  2. [Method] Method section on frequency cluster aggregation: no analysis (visualization, alignment metrics, or cross-subject consistency check) is provided to demonstrate that the learned clusters correspond to established EEG rhythms rather than dataset-specific frequency distributions; this directly affects the claim that the module captures generalizable stress-related patterns.
  3. [Method] Intra-inter slice attention module description: the manuscript does not report an ablation isolating the contribution of the attention mechanism versus the Riemannian spectral representation, leaving open whether the reported gains arise from the proposed components or from other implementation choices.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'up to 82.78%' is used without indicating the dataset or the corresponding mean performance across the three datasets.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which help strengthen the manuscript. We address each major point below and will revise accordingly to improve clarity on validation protocols, add supporting analyses for the frequency module, and include ablations for the attention component.

read point-by-point responses

  1. Referee: [Experiments] Experiments section: performance figures (including the 82.78% peak) are stated without error bars, statistical tests, or explicit description of the cross-validation protocol (e.g., leave-one-subject-out or stratified subject splits), which is load-bearing for the central claim of consistent outperformance and generalizability.

    Authors: We agree this information is essential for substantiating the cross-subject claims. The original experiments used leave-one-subject-out cross-validation on all three datasets with subject-independent splits; we will explicitly state this protocol, report mean balanced accuracy with standard deviation error bars across folds, and add statistical comparisons (paired t-tests) against baselines in the revised Experiments section. revision: yes

  2. Referee: [Method] Method section on frequency cluster aggregation: no analysis (visualization, alignment metrics, or cross-subject consistency check) is provided to demonstrate that the learned clusters correspond to established EEG rhythms rather than dataset-specific frequency distributions; this directly affects the claim that the module captures generalizable stress-related patterns.

    Authors: We acknowledge that interpretability evidence would better support the generalizability claim. In revision we will add visualizations of the learned cluster centroids overlaid on canonical EEG bands (delta/theta/alpha/beta/gamma) plus quantitative alignment metrics (e.g., frequency-range overlap ratios) computed across subjects; this directly addresses the concern while preserving the data-driven nature of the aggregation. revision: yes

  3. Referee: [Method] Intra-inter slice attention module description: the manuscript does not report an ablation isolating the contribution of the attention mechanism versus the Riemannian spectral representation, leaving open whether the reported gains arise from the proposed components or from other implementation choices.

    Authors: We agree an ablation is needed to isolate component contributions. The revised manuscript will include a dedicated ablation study comparing the full I²RiMA against (i) the Riemannian spectral representation alone and (ii) the representation plus intra-inter attention, quantifying the incremental gains on all three datasets. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical validation of proposed architecture

full rationale

The paper proposes an architecture (frequency-specific SPD covariance, cluster aggregation, intra-inter attention) and reports experimental accuracies on three datasets against baselines. No derivation chain, first-principles result, or prediction is claimed that reduces by the paper's own equations to its inputs. No self-citation load-bearing steps, fitted-input-as-prediction, or ansatz smuggling appear in the provided text. The central claim is empirical performance, which is self-contained against external benchmarks and does not trigger any of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on empirical performance of a new architecture whose components are introduced without first-principles derivation; effectiveness is demonstrated on three datasets whose representativeness is assumed.

free parameters (2)
  • number of frequency clusters
    Data-driven selection but the exact count and selection criterion function as a tunable hyperparameter.
  • attention module dimensions and heads
    Learned during training; specific architecture sizes are free parameters of the model.
axioms (2)
  • domain assumption Spatial covariance matrices computed independently at each frequency point preserve channel-wise geometry together with frequency-specific cues for stress.
    Invoked when constructing per-frequency SPD matrices and mapping to tangent space.
  • domain assumption Frequency cluster aggregation reduces redundancy while retaining informative spectral components aligned with EEG rhythms.
    Used to justify the aggregation step before attention.

pith-pipeline@v0.9.1-grok · 5740 in / 1398 out tokens · 30091 ms · 2026-07-03T21:43:05.853466+00:00 · methodology

0 comments
read the original abstract

Cross-subject EEG stress detection remains challenging because discriminative stress-related patterns are both subject-dependent and frequency-specific. Conventional Riemannian methods model spatial covariance mainly in the time domain, overlooking neural oscillations that are critical for high-level cognitive state decoding, while standard temporal tokenization often fragments inter-slice temporal coherence. To address these limitations, we propose \method{}, an Intra-Inter Riemannian Manifold Attention Network for EEG-based stress detection. \method{} constructs spatial covariance matrices independently at each frequency point and maps them to the SPD tangent space, preserving channel-wise geometry together with frequency-specific discriminative cues. It further introduces frequency cluster aggregation to select informative spectral components and reduce redundancy by forming compact, data-driven frequency clusters aligned with EEG rhythms. Finally, an intra-inter slice attention module adaptively integrates local slice-level spectral dynamics and global temporal context across EEG sequences. Experiments on three datasets show that \method{} consistently outperforms five state-of-the-art baselines, achieving up to 82.78\% balanced accuracy while remaining efficient with only 1.60M parameters and 31.95M FLOPs.

Figures

Figures reproduced from arXiv: 2607.01279 by Cheng He, Jinhong Ding, Jinming Ma, Kunyu Peng, Likun Xia, Shangen Han.

Figure 1
Figure 1. Figure 1: Overview of I2RiMA. The pipeline comprises four stages: Preprocessing and FFT; USAA Module to perform frequency cluster aggregation via K-Means and intra-inter slice attention fusion; RMFE Module to perform frequency-wise covariance construction and Log-Euclidean tangent-space mapping; Classification for mental stress detection. The Unsupervised Slice Attention Aggregation (USAA) module then processes thes… view at source ↗
Figure 2
Figure 2. Figure 2: Model Performance Comparison) 5.2 Ablation Study To assess each component of I2RiMA, we design four ablation variants: baseline without Riemannian modeling or inter-slice fusion, R-I2RiMA using only the Riemannian module (m=1), I-I2RiMA using only inter-slice fusion, and the full I2RiMA. Results are reported in [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Channel Importance Topographic Maps of I [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Topographical Maps of Channel Discriminability Across Temporal Windows (Welch [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Comparison across Datasets and Methods in Spectrum domain. [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison across Datasets and Methods in Spectrum domain. [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Marginal Effect Comparison across Three Datasets [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages

  1. [1]

    Depression recognition based on the reconstruction of phase space of eeg signals and geometrical features.Applied Acoustics, 179:108078, 2021

    Hesam Akbari, Muhammad Tariq Sadiq, Ateeq Ur Rehman, Mahdieh Ghazvini, Rizwan Ali Naqvi, Malih Payan, Hourieh Bagheri, and Hamed Bagheri. Depression recognition based on the reconstruction of phase space of eeg signals and geometrical features.Applied Acoustics, 179:108078, 2021

    work page 2021

  2. [2]

    A wearable eeg instrument for real-time frontal asymmetry monitoring in worker stress analysis

    Pasquale Arpaia, Nicola Moccaldi, Roberto Prevete, Isabella Sannino, and Annarita Tedesco. A wearable eeg instrument for real-time frontal asymmetry monitoring in worker stress analysis. IEEE Transactions on Instrumentation and Measurement, 69(10):8335–8343, 2020

    work page 2020

  3. [3]

    Riemannian geometry applied to bci classification

    Alexandre Barachant, Stéphane Bonnet, Marco Congedo, and Christian Jutten. Riemannian geometry applied to bci classification. InInternational Conference on Latent Variable Analysis and Signal Separation, pages 629–636. Springer, 2010

    work page 2010

  4. [4]

    Multiclass brain– computer interface classification by riemannian geometry.IEEE Transactions on Biomedical Engineering, 59(4):920–928, 2011

    Alexandre Barachant, Stéphane Bonnet, Marco Congedo, and Christian Jutten. Multiclass brain– computer interface classification by riemannian geometry.IEEE Transactions on Biomedical Engineering, 59(4):920–928, 2011

    work page 2011

  5. [5]

    Electroencephalography (eeg): An introductory text and atlas of normal and abnormal findings in adults, children, and infants

    Jeffrey W Britton, Lauren C Frey, Jennifer L Hopp, Pearce Korb, Mohamad Z Koubeissi, William E Lievens, Elia M Pestana-Knight, and Erk K St Louis. Electroencephalography (eeg): An introductory text and atlas of normal and abnormal findings in adults, children, and infants. 2016

    work page 2016

  6. [6]

    The origin of extracellular fields and currents—eeg, ecog, lfp and spikes.Nature Reviews Neuroscience, 13(6):407–420, 2012

    György Buzsáki, Costas A Anastassiou, and Christof Koch. The origin of extracellular fields and currents—eeg, ecog, lfp and spikes.Nature Reviews Neuroscience, 13(6):407–420, 2012

    work page 2012

  7. [7]

    Study on feature selection methods for depression detection using three-electrode eeg data.Interdisciplinary Sciences: Computational Life Sciences, 10(3):558–565, 2018

    Hanshu Cai, Yunfei Chen, Jiashuo Han, Xiangzi Zhang, and Bin Hu. Study on feature selection methods for depression detection using three-electrode eeg data.Interdisciplinary Sciences: Computational Life Sciences, 10(3):558–565, 2018

    work page 2018

  8. [8]

    Riemannian geometry for eeg- based brain-computer interfaces; a primer and a review.Brain-Computer Interfaces, 4(3):155– 174, 2017

    Marco Congedo, Alexandre Barachant, and Rajendra Bhatia. Riemannian geometry for eeg- based brain-computer interfaces; a primer and a review.Brain-Computer Interfaces, 4(3):155– 174, 2017

    work page 2017

  9. [9]

    Eeg and meg: relevance to neuroscience.Neuron, 80(5):1112–1128, 2013

    Fernando Lopes da Silva. Eeg and meg: relevance to neuroscience.Neuron, 80(5):1112–1128, 2013

    work page 2013

  10. [10]

    Emotion, plasticity, context, and regulation: perspectives from affective neuroscience.Psychological Bulletin, 126(6):890, 2000

    Richard J Davidson, Daren C Jackson, and Ned H Kalin. Emotion, plasticity, context, and regulation: perspectives from affective neuroscience.Psychological Bulletin, 126(6):890, 2000

    work page 2000

  11. [11]

    Report on national mental health development in china (2021–2022).Social Sciences Academic Press, pages 70–99, 2023

    Xiaolan Fu, Kai Zhang, X Chen, and Z Chen. Report on national mental health development in china (2021–2022).Social Sciences Academic Press, pages 70–99, 2023

    work page 2021

  12. [12]

    Review on psychological stress detection using biosignals

    Giorgos Giannakakis, Dimitris Grigoriadis, Katerina Giannakaki, Olympia Simantiraki, Alexan- dros Roniotis, and Manolis Tsiknakis. Review on psychological stress detection using biosignals. IEEE Transactions on Affective Computing, 13(1):440–460, 2019. 10

    work page 2019

  13. [13]

    Context sensitivity of eeg-based workload classification under different affective valence.IEEE Transactions on Affective Computing, 11(2):327–334, 2017

    Sebastian Grissmann, Martin Spüler, Josef Faller, Tanja Krumpe, Thorsten O Zander, Augustin Kelava, Christian Scharinger, and Peter Gerjets. Context sensitivity of eeg-based workload classification under different affective valence.IEEE Transactions on Affective Computing, 11(2):327–334, 2017

    work page 2017

  14. [14]

    Dicha: Dynamic and interpretable convolution via hierarchical attention for robust eeg classification.Biomedical Signal Processing and Control, 120:110091, 2026

    Cheng He, Ziheng Guo, Yehan Xu, Xiaoxu Zhao, Wanying Yang, Xiaoyu Wang, Jinhong Ding, Ming Ma, Jingxu Chen, Shuangjiang Zhou, and Likun Xia. Dicha: Dynamic and interpretable convolution via hierarchical attention for robust eeg classification.Biomedical Signal Processing and Control, 120:110091, 2026

    work page 2026

  15. [15]

    A correlation manifold self-attention network for eeg decoding

    Chen Hu, Rui Wang, Xiaoning Song, Tao Zhou, Xiao-Jun Wu, Nicu Sebe, and Ziheng Chen. A correlation manifold self-attention network for eeg decoding. InProceedings of the Thirty- Fourth International Joint Conference on Artificial Intelligence, pages 5372–5380, 2025

    work page 2025

  16. [16]

    Transfer learning in brain-computer interfaces.IEEE Computational Intelligence Magazine, 11(1):20–31, 2016

    Vinay Jayaram, Morteza Alamgir, Yasemin Altun, Bernhard Scholkopf, and Moritz Grosse- Wentrup. Transfer learning in brain-computer interfaces.IEEE Computational Intelligence Magazine, 11(1):20–31, 2016

    work page 2016

  17. [17]

    Large Brain Model for Learning Generic Representations with Tremendous EEG Data in BCI

    Wei-Bang Jiang, Li-Ming Zhao, and Bao-Liang Lu. Large brain model for learning generic representations with tremendous eeg data in bci.arXiv preprint arXiv:2405.18765, 2024

    work page Pith review arXiv 2024

  18. [18]

    Allostatic load biomarkers of chronic stress and impact on health and cognition.Neuroscience & Biobehavioral Reviews, 35(1):2–16, 2010

    Robert-Paul Juster, Bruce S McEwen, and Sonia J Lupien. Allostatic load biomarkers of chronic stress and impact on health and cognition.Neuroscience & Biobehavioral Reviews, 35(1):2–16, 2010

    work page 2010

  19. [19]

    Eegnet: a compact convolutional neural network for eeg-based brain–computer interfaces.Journal of Neural Engineering, 15(5):056013, 2018

    Vernon J Lawhern, Amelia J Solon, Nicholas R Waytowich, Stephen M Gordon, Chou P Hung, and Brent J Lance. Eegnet: a compact convolutional neural network for eeg-based brain–computer interfaces.Journal of Neural Engineering, 15(5):056013, 2018

    work page 2018

  20. [20]

    Neurobolt: Resting-state eeg-to-fmri synthesis with multi-dimensional feature mapping.Advances in Neural Information Processing Systems, 37:23378–23405, 2024

    Yamin Li, Ange Lou, Ziyuan Xu, Shengchao Zhang, Shiyu Wang, Dario J Englot, Soheil Kolouri, Daniel Moyer, Roza G Bayrak, and Catie Chang. Neurobolt: Resting-state eeg-to-fmri synthesis with multi-dimensional feature mapping.Advances in Neural Information Processing Systems, 37:23378–23405, 2024

    work page 2024

  21. [21]

    A review of classification algorithms for eeg-based brain– computer interfaces: a 10 year update.Journal of Neural Engineering, 15(3):031005, 2018

    Fabien Lotte, Laurent Bougrain, Andrzej Cichocki, Maureen Clerc, Marco Congedo, Alain Rakotomamonjy, and Florian Yger. A review of classification algorithms for eeg-based brain– computer interfaces: a 10 year update.Journal of Neural Engineering, 15(3):031005, 2018

    work page 2018

  22. [22]

    Enhanced classification of motor imagery eeg signals using spatio-temporal representa- tions.Information Sciences, 714:122221, 2025

    Renjie Lv, Wenwen Chang, Guanghui Yan, Muhammad Tariq Sadiq, Wenchao Nie, and Lei Zheng. Enhanced classification of motor imagery eeg signals using spatio-temporal representa- tions.Information Sciences, 714:122221, 2025

    work page 2025

  23. [23]

    Secondary depression in severe anxiety disorders: a population-based cohort study in denmark.The Lancet Psychiatry, 2(6):515–523, 2015

    Sandra M Meier, Liselotte Petersen, Manuel Mattheisen, Ole Mors, Preben B Mortensen, and Thomas M Laursen. Secondary depression in severe anxiety disorders: a population-based cohort study in denmark.The Lancet Psychiatry, 2(6):515–523, 2015

    work page 2015

  24. [24]

    Adjust: An automatic eeg artifact detector based on the joint use of spatial and temporal features.Psychophysiology, 48(2):229–240, 2011

    Andrea Mognon, Jorge Jovicich, Lorenzo Bruzzone, and Marco Buiatti. Adjust: An automatic eeg artifact detector based on the joint use of spatial and temporal features.Psychophysiology, 48(2):229–240, 2011

    work page 2011

  25. [25]

    Matt: A manifold attention network for eeg decoding.Advances in Neural Information Processing Systems, 35:31116–31129, 2022

    Yue-Ting Pan, Jing-Lun Chou, and Chun-Shu Wei. Matt: A manifold attention network for eeg decoding.Advances in Neural Information Processing Systems, 35:31116–31129, 2022

    work page 2022

  26. [26]

    Eeg-based multiclass workload identification using feature fusion and selection.IEEE Transactions on Instrumentation and Measurement, 70:1–8, 2020

    Zian Pei, Hongtao Wang, Anastasios Bezerianos, and Junhua Li. Eeg-based multiclass workload identification using feature fusion and selection.IEEE Transactions on Instrumentation and Measurement, 70:1–8, 2020

    work page 2020

  27. [27]

    The global prevalence of common mental disorders: a systematic review and meta-analysis 1980–2013.International Journal of Epidemiology, 43(2):476–493, 2014

    Zachary Steel, Claire Marnane, Changiz Iranpour, Tien Chey, John W Jackson, Vikram Patel, and Derrick Silove. The global prevalence of common mental disorders: a systematic review and meta-analysis 1980–2013.International Journal of Epidemiology, 43(2):476–493, 2014

    work page 1980

  28. [28]

    Neural regulation of endocrine and autonomic stress responses.Nature Reviews Neuroscience, 10(6):397–409, 2009

    Yvonne M Ulrich-Lai and James P Herman. Neural regulation of endocrine and autonomic stress responses.Nature Reviews Neuroscience, 10(6):397–409, 2009. 11

    work page 2009

  29. [29]

    Attention is all you need.Advances in Neural Information Processing Systems, 30, 2017

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need.Advances in Neural Information Processing Systems, 30, 2017

    work page 2017

  30. [30]

    A grassmannian manifold self-attention network for signal classification

    Rui Wang, Chen Hu, Ziheng Chen, Xiao-Jun Wu, and Xiaoning Song. A grassmannian manifold self-attention network for signal classification. InProceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, pages 5099–5107, 2024

    work page 2024

  31. [31]

    Analysing heart- brain coupling: A correlation study of bcg, eeg, and ecg signals

    Haoyu Wu, Jingzhou Xu, Xiaopeng Lu, Qing Liu, Jianjun Chen, and Jun Qi. Analysing heart- brain coupling: A correlation study of bcg, eeg, and ecg signals. In2025 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pages 6133–6140. IEEE, 2025

    work page 2025

  32. [32]

    Likun Xia, Yuan Feng, Ziheng Guo, Jinhong Ding, Yanlei Li, Yifan Li, Ming Ma, Guoxi Gan, Yehan Xu, Jingyu Luo, et al. Mulhita: A novel multiclass classification framework with multibranch lstm and hierarchical temporal attention for early detection of mental stress.IEEE Transactions on Neural Networks and Learning Systems, 34(12):9657–9670, 2022

    work page 2022

  33. [33]

    A physiological signal-based method for early mental-stress detection.Biomedical Signal Processing and Control, 46:18–32, 2018

    Likun Xia, Aamir Saeed Malik, and Ahmad Rauf Subhani. A physiological signal-based method for early mental-stress detection.Biomedical Signal Processing and Control, 46:18–32, 2018

    work page 2018

  34. [34]

    Biot: Biosignal transformer for cross-data learning in the wild.Advances in Neural Information Processing Systems, 36:78240–78260, 2023

    Chaoqi Yang, M Westover, and Jimeng Sun. Biot: Biosignal transformer for cross-data learning in the wild.Advances in Neural Information Processing Systems, 36:78240–78260, 2023

    work page 2023

  35. [35]

    Riemannian approaches in brain-computer interfaces: a review.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 25(10):1753–1762, 2016

    Florian Yger, Maxime Berar, and Fabien Lotte. Riemannian approaches in brain-computer interfaces: a review.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 25(10):1753–1762, 2016

    work page 2016

  36. [36]

    Matteo Zanetti, Teruhiro Mizumoto, Luca Faes, Alberto Fornaser, Mariolino De Cecco, Luca Maule, Martina Valente, and Giandomenico Nollo. Multilevel assessment of mental stress via network physiology paradigm using consumer wearable devices.Journal of Ambient Intelligence and Humanized Computing, 12(4):4409–4418, 2021

    work page 2021

  37. [37]

    Stress-induced effects in resting eeg spectra predict the performance of ssvep-based bci.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 28(8):1771–1780, 2020

    Hao-Yan Zhang, Cory E Stevenson, Tzyy-Ping Jung, and Li-Wei Ko. Stress-induced effects in resting eeg spectra predict the performance of ssvep-based bci.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 28(8):1771–1780, 2020

    work page 2020

  38. [38]

    Pengbo Zhang, Xue Wang, Weihang Zhang, and Junfeng Chen. Learning spatial–spectral– temporal eeg features with recurrent 3d convolutional neural networks for cross-task mental workload assessment.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 27(1):31–42, 2018. Appendix A Overview This appendix provides supplementary material organize...

    work page 2018