Pith. sign in

REVIEW 3 major objections 4 minor 52 references

MR-EEGWaveNet: Multiresolutional EEGWaveNet for Seizure Detection from Long EEG Recordings

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read MR-EEGWaveNet roughly doubles seizure-detection F1 on two long scalp-EEG datasets by pooling features from whole segments and sub-segments and filtering outputs with an anomaly-score threshold.

desk verdict The multiresolution idea is real, but the abstract's headline gains come from a post-processing rule that, as written, cannot depend on the architecture. read the letter →

arxiv 2505.17972 v2 pith:QXDFQPOC submitted 2025-05-23 cs.CV cs.LG

Kazi Mahmudul Hassan , Xuyang Zhao , Hidenori Sugano , Toshihisa Tanaka This is my paper · ORCID
classification cs.CVcs.LG
keywords seizuredetectionelectroencephalogrammultiresolutionaldeeplearningEEGWaveNetanomalyscorepost-classificationECODleave-one-subject-outvalidationscalpEEGartifacts
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that a single-resolution seizure detector can be improved by analyzing each EEG segment at several time resolutions at once. Its model, MR-EEGWaveNet, feeds a long segment and its shorter sub-segments through the same convolutional feature extractor, concatenates the resulting features, and then applies an anomaly-score threshold to cut false alarms. On the public Siena dataset the F1 score rises from 0.177 to 0.336, and on the private Juntendo dataset from 0.327 to 0.488, with precision gains of about 16 and 21 percentage points. The improvements matter because scalp EEG is noisy and imbalanced, and a detector that reduces false positives without losing many true seizure segments is the kind of system clinicians could actually use.

What carries the argument

The central object is the multiresolutional feature vector formed by concatenating the 32-dimensional outputs of the shared convolution-plus-feature-extraction modules for a segment of length $W$ and for each of its sub-segments of lengths in $D$, giving $K = \sum_{d \in D} F \lfloor W/d \rfloor$ features that the predictor module maps to a class. The second mechanism is the post-classification rule: each segment's ECOD anomaly score is compared with the mean anomaly score of all segments in the recording, and segments below the mean are relabeled nonseizure. Together they let the model keep the statistical context of a long window while retaining the fine temporal detail of short windows, and then suppress scattered false positives.

What would settle it

Re-run the LOSO experiments with the ECOD threshold computed from training-fold segments only (or fixed at a percentile), and check whether the F1 gains from 0.177 to 0.336 on Siena and from 0.327 to 0.488 on Juntendo survive; if they shrink or vanish, the reported advantage comes from the per-recording threshold rather than from multiresolution features.

Watch

Extended reading notes

Core claim

The central claim is that multiresolutional feature extraction plus anomaly-score post-processing makes EEGWaveNet markedly better at separating seizures from background EEG and artifacts in long recordings. For a 10-second segment, the model also looks at five 2-second sub-segments, computes a 32-dimensional feature vector for each, concatenates them, and classifies with a small fully connected network. The mean anomaly score of the test recording, computed by ECOD on the raw segments, is then used as a threshold to relabel low-scoring positives as nonseizure. The stated outcome is F1 rising from 0.177 to 0.336 on Siena and from 0.327 to 0.488 on Juntendo, with specificity rising to about 95 percent on both datasets while recall stays near 78-80 percent. Ablations attribute part of the gain to the feature extraction module and part to the 10-second context stream, and the full model also beats a frozen foundation-model baseline on the same leave-one-subject-out protocol.

Load-bearing premise

The post-classification threshold is the mean anomaly score of the very test recording being evaluated, and the paper assumes this per-recording threshold is a valid, unbiased part of the evaluation rather than test-set leakage or an artifact of threshold tuning on the same data.

Editorial extensions

If this is right

  • Window length alone moves performance: longer windows improve capture of seizure regularity but lose temporal detail, and the multiresolution design combines both effects.
  • The anomaly-score post-processing step improves precision and specificity across all tested single- and multi-resolution models, at a small recall cost.
  • In leave-one-subject-out evaluation, MR-EEGWaveNet-2 [10 s, 2 s] gives the best balance, and removing either the feature-extraction module or the 10-second stream degrades F1 and AUC.
  • On the Siena dataset the model detects about 94 percent of annotated seizure events, outperforming the closest LOSO baseline in event-based detection by about 13 points while trailing in specificity by about 4 points.
  • A task-specific end-to-end model can outperform a frozen large pretrained EEG foundation model plus gradient-boosted classifier on this seizure-detection benchmark.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The mean-anomaly threshold is computed per test recording; in a prospective clinical deployment the full recording would not be available at decision time, so a threshold trained on the training distribution or fixed at a quantile would be the honest online variant.
  • Because the multiresolution gain is measured against EEGWaveNet with different segment lengths but not against a single-resolution model matched in parameter count or compute, part of the improvement could in principle come from added capacity; a parameter-matched ablation would isolate the multiresolution mechanism.
  • The same ECOD-based post-classification rule could be applied to other segment-level EEG classifiers; testing whether it generalizes across datasets and montages would clarify whether the reported gains are architecture-specific.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes MR-EEGWaveNet, a multiresolution extension of EEGWaveNet for seizure detection from long multichannel EEG recordings. The architecture adds sub-segment feature extraction at multiple resolutions, and the authors introduce an anomaly-score-based post-classification rule intended to reduce false positives. The model is evaluated with leave-one-subject-out cross-validation on the public Siena dataset and a private Juntendo dataset, with comparisons to EEGWaveNet, ablated variants, and a pretrained foundation model. The abstract's headline improvement is the post-processed F1 increase from 0.177 to 0.336 on Siena and 0.327 to 0.488 on Juntendo.

Significance. If the results were reproducible, the contribution would be a useful incremental improvement: the raw comparisons in Tables 4 and 5 do show that the multiresolution architecture improves precision and F1 over EEGWaveNet, and the use of LOSO evaluation on two datasets, including a private clinical dataset, is a strength. The paper also provides ablation studies and a code repository. However, the central post-classification mechanism, which supplies the abstract's headline numbers, is internally inconsistent as written, and the threshold choice uses test-recording information without independent validation. These issues currently prevent the reported post-processed gains from being evaluated or reproduced.

major comments (3)
  1. [Section 4.2, Eqs. (1)-(2); Section 5; Tables 6-7] The post-classification rule as written cannot produce the reported results. In Eq. (2), the final label Li depends only on the ECOD anomaly score ai and the recording-level mean threshold, and Section 5 states that ECOD is applied directly to raw z-scored flattened EEG segments, not to classifier outputs. Therefore, for a fixed test segmentation and a fixed window length, every model would receive identical post-processed labels. Yet Tables 6 and 7 report model-dependent post-processed performance for models sharing the same window length; for example, on Siena 10-s segments, EEGWaveNet-3 has post-processed F1 = 0.177 while MR-EEGWaveNet-2 has post-processed F1 = 0.336. Because the abstract's central quantitative claim is taken from these post-processed rows, the actual post-classification mechanism must be disclosed and the experiments re-reported with a self-consistent rule.
  2. [Section 4.2, Eq. (1); Section 7.2] The threshold µa is computed on the full test recording used for evaluation, and Section 7.2 explains that the mean was chosen because experimental results supported it. This means test-set information enters the decision rule and no independent threshold validation is reported. The resulting post-processed precision and F1 numbers are therefore not a clean evaluation of the proposed method; the threshold selection should be performed on training or validation data, or an explicit offline-analysis justification with a separate threshold-fixing procedure should be given.
  3. [Section 7.7, Table 10] The event-based comparison with Sigsgaard et al. reports only recall and specificity, and the event-detection definition is introduced only later in the same section. The claim that the proposed method outperforms the prior work by approximately 13% in seizure detection is not supported by a matched evaluation protocol, since the compared method uses a different evaluation scheme and different event criteria. The comparison should be made on identical segment- or event-level metrics, or clearly labeled as not directly comparable.
minor comments (4)
  1. [Section 5] The term 'post-classification processing' is misleading if ECOD is applied directly to raw EEG segments rather than to classifier outputs; the terminology should be aligned with the actual processing chain, especially given the inconsistency flagged in the major comments.
  2. [Tables 6 and 7] AUC values are omitted for all post-processed rows, making it difficult to assess whether the post-processing step changes ranking quality or only threshold-based metrics; reporting AUC for these rows would strengthen the analysis.
  3. [Section 7.5] The Wilcoxon signed rank test is described only by the resulting p-value; the manuscript should state the test statistic, the number of paired patients, and which patient-wise metric (F1, precision, or AUC) was used for the comparison.
  4. [Figures 2a and 2b] The figure labels contain formatting artifacts, such as 'C) Predictor Module' and 'B) Feature Extraction Module§Fully Connected...', which should be cleaned up for readability.

Circularity Check

1 steps flagged · score 3.0 of 10

Headline gains come from the post-classification rows, whose threshold is computed on the same test recordings and whose written rule cannot produce the reported model-dependent F1 differences.

  1. fitted input called prediction [Section 4.2 (Eqs. 1-2) and Section 7.2]
    "µa = 1/M Σ M i=1 ai (1) Li = ( 1, seizure, if ai > µa 0, nonseizure. (2) where µa is the mean of ai, M is the total number of segments... Based on this observation, the mean anomaly score was chosen as the threshold, with experimental results supporting its effectiveness."

    The post-classification threshold μ_a is computed from the very test-recording anomaly scores that Eq. 2 then thresholds, and the choice of the mean is justified by the experimental outcomes it is used to produce. The post-processed F1/precision figures in Tables 6-7 are therefore not held-out predictions of a fixed rule; they are the result of a threshold adaptively chosen on the test set. Moreover, Section 5 states ECOD is applied directly to raw flattened z-scored EEG, so under Eq. 2 the labels for identical 10-s test segments would be the same for EEGWaveNet-3 and MR-EEGWaveNet-2, yet Table 6 reports post-processed F1 = 0.177 vs 0.336.

full rationale

The core architectural claim, that multiresolution inputs improve over single-window EEGWaveNet, is independently supported by the raw (non-post-processed) rows: e.g., Siena F1 0.319 vs 0.161 and Juntendo F1 0.474 vs 0.274 for MR-EEGWaveNet-2 vs EEGWaveNet-3. That part is not circular. The circularity is confined to the post-classification processing that supplies the abstract's headline numbers (0.177 to 0.336 and 0.327 to 0.488). Equations 1-2 compute the threshold from the test recording and then evaluate on that same recording, and Section 7.2 admits the threshold was chosen because experimental results supported it, making the post-processed performance an in-sample fit rather than a prediction. Additionally, as written, Eq. 2 makes the post-processed labels depend only on ECOD scores of raw EEG segments and the recording-level mean, so the different post-processed F1 values across same-window models in Tables 6-7 cannot be reproduced from the described rule. I therefore rate the circularity moderate (3/10): enough to undermine the headline post-processing gains, but not the entire architecture comparison, which has independent content in the raw results.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. The main assumptions are domain-specific beliefs about EEG regularity and the behavior of ECOD scores, plus a methodological assumption about the oversampling scheme not causing leakage. The free parameters are architecture and threshold choices made after inspecting test performance.

free parameters (5)
  • window length W = 2s, 5s, 10s (tested)
    Segmentation window length is a hyperparameter selected by comparing model variants; performance varies with W, so the choice of 10s for the final model is made after seeing test results.
  • multiresolution parameter list D = [5s, 2.5s], [10s, 2s], [10s, 5s, 2s]
    The number and lengths of sub-segments are chosen by the authors; MR-EEGWaveNet-2 was selected as the primary model after comparing test performance, a form of post-hoc model selection.
  • class weights = 0.75 (nonseizure), 1.5 (seizure)
    Fixed hyperparameters, not fitted but chosen by hand to counter class imbalance.
  • learning rate, batch size, epochs, patience = 1e-3, 32, 300, 30
    Standard hyperparameters that were fixed for all experiments without a reported search.
  • ECOD threshold rule = mean anomaly score
    The threshold is defined as the mean of anomaly scores over each test recording and was chosen based on experimental results (Section 7.2), so it is an ad hoc rule not derived from theory.
assumptions (4)
  • domain assumption Seizure EEG exhibits higher signal regularity than nonseizure EEG.
    Invoked in Discussion 7.1 to motivate longer windows and multiresolution analysis; based on prior literature [22, 1].
  • domain assumption ECOD anomaly scores on raw flattened EEG segments separate seizure and artifact segments from normal background.
    The post-classification processing assumes that anomalous EEG segments correspond to seizures/artifacts and that the mean anomaly score is a valid threshold (Section 4.2).
  • domain assumption The 80% overlapping oversampling of seizure segments does not cause harmful train/validation leakage.
    Training procedure oversamples seizure segments with 80% overlap, then randomly splits into 70% train / 30% validation; overlapping segments from the same seizure can appear in both, potentially biasing early stopping (Section 5).
  • standard math ECOD implementation from PyOD works as documented.
    The anomaly detection method is used as a black box from the PyOD library [47, 48].

how reviews work

0 comments
Cite this review

Pith. "Pith review of MR-EEGWaveNet: Multiresolutional EEGWaveNet for Seizure Detection from Long EEG Recordings." pith.science (2026). https://pith.science/paper/QXDFQPOC

@misc{pith2026250517972,
  author       = {Pith},
  title        = {Pith review of: MR-EEGWaveNet: Multiresolutional EEGWaveNet for Seizure Detection from Long EEG Recordings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QXDFQPOC}},
  note         = {Machine review of arXiv:2505.17972}
}
read the original abstract

Feature engineering for generalized seizure detection models remains a significant challenge. Recently proposed models show variable performance depending on the training data and remain ineffective at accurately distinguishing artifacts from seizure data. In this study, we propose a novel end-to-end model, "Multiresolutional EEGWaveNet (MR-EEGWaveNet)," which efficiently distinguishes seizure events from background electroencephalogram (EEG) and artifacts/noise by capturing both temporal dependencies across different time frames and spatial relationships between channels. The model has three modules: convolution, feature extraction, and predictor. The convolution module extracts features through depth-wise and spatio-temporal convolution. The feature extraction module individually reduces the feature dimension extracted from EEG segments and their sub-segments. Subsequently, the extracted features are concatenated into a single vector for classification using a fully connected classifier called the predictor module. In addition, an anomaly score-based post-classification processing technique is introduced to reduce the false-positive rates of the model. Experimental results are reported and analyzed using different parameter settings and datasets (Siena (public) and Juntendo (private)). The proposed MR-EEGWaveNet significantly outperformed the conventional non-multiresolution approach, improving the F1 scores from 0.177 to 0.336 on Siena and 0.327 to 0.488 on Juntendo, with precision gains of 15.9% and 20.62%, respectively.

Figures

Figures reproduced from arXiv: 2505.17972 by the authors.

Figure 1
Figure 1. 10-seconds long (a) Nonseizure, (b) Seizure EEG segments from Juntendo dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Classification process of MR-EEGWaveNet model (a) Convolution module, and (b) [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Illustration of Leave-One-Subject-Out (LOSO) cross-validation scheme [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Outcome of the MR-EEGWaveNet-2 model on a 1-hour EEG recording after post [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]
Figure 5
Figure 5. Figure 5: t-SNE visualization of features extracted from EEG recordings using EEGWaveNet [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 52 canonical work pages

  1. [1]

    Epileptic seizure detection from EEG signals using multiband features with feedforward neural network,

    K. M. Hassan, M. R. Islam, T. Tanaka, and M. K. I. Molla, “Epileptic seizure detection from EEG signals using multiband features with feedforward neural network,” in 2019 International Conference on Cyberworlds (CW) , pp. 231–238, 2019

    work page 2019

  2. [2]

    ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology,

    I. E. Scheffer, S. Berkovic, G. Capovilla, M. B. Connolly, J. French, L. Guilhoto, E. Hirsch, S. Jain, G. W. Mathern, S. L. Mosh´ e, D. R. Nordli, E. Perucca, T. Tomson, S. Wiebe, Y. Zhang, and S. M. Zuberi, “ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology,” Epilepsia, vol. 58, pp. 512–521, Apr. 2017

    work page 2017

  3. [3]

    A review of seizures and epilepsy following traumatic brain injury,

    S. Fordington and M. Manford, “A review of seizures and epilepsy following traumatic brain injury,” Journal of Neurology , vol. 267, pp. 3105–3111, Oct. 2020

    work page 2020

  4. [4]

    Seizure detection using scalp-EEG,

    C. Baumgartner and J. P. Koren, “Seizure detection using scalp-EEG,” Epilepsia, vol. 59, no. S1, pp. 14–22, 2018

    work page 2018

  5. [5]

    Clinical translation of machine learning algorithms for seizure detection in scalp electroencephalography: systematic review,

    N. Moutonnet, S. White, B. P. Campbell, S. Sanei, T. Tanaka, H. Ji, D. Mandic, and G. Scott, “Clinical translation of machine learning algorithms for seizure detection in scalp electroencephalography: systematic review,” 2024

    work page 2024

  6. [6]

    A review of epilepsy detection and prediction methods based on EEG signal processing and deep learning,

    X. Zhang, X. Zhang, Q. Huang, and F. Chen, “A review of epilepsy detection and prediction methods based on EEG signal processing and deep learning,” Frontiers in Neuroscience, vol. Volume 18 - 2024, 2024

    work page 2024

  7. [7]

    LightSeizureNet: A lightweight deep learning model for real-time epileptic seizure detection,

    S. Qiu, W. Wang, and H. Jiao, “LightSeizureNet: A lightweight deep learning model for real-time epileptic seizure detection,” IEEE Journal of Biomedical and Health Informat- ics, vol. 27, pp. 1845–1856, Apr. 2023

    work page 2023

  8. [8]

    Deep learning based epileptic seizure detection with EEG data,

    S. Poorani and P. Balasubramanie, “Deep learning based epileptic seizure detection with EEG data,” International Journal of System Assurance Engineering and Management , Jan. 2023

    work page 2023

Show all 52 references
  1. [9]

    XAI4EEG: spectral and spatio-temporal explanation of deep learning-based seizure detection in EEG time series,

    D. Raab, A. Theissler, and M. Spiliopoulou, “XAI4EEG: spectral and spatio-temporal explanation of deep learning-based seizure detection in EEG time series,” Neural Com- puting and Applications , vol. 35, pp. 10051–10068, May 2023

    2023

  2. [10]

    Deep-learning-based seizure detection and prediction from electroencephalography sig- nals,

    F. E. Ibrahim, H. M. Emara, W. El-Shafai, M. Elwekeil, M. Rihan, I. M. Eldokany, T. E. Taha, A. S. El-Fishawy, E. M. El-Rabaie, E. Abdellatef, and F. E. Abd El-Samie, “Deep-learning-based seizure detection and prediction from electroencephalography sig- nals,” International Jo...

    2022

  3. [11]

    A scheme combining feature fusion and hybrid deep learning models for epileptic seizure detection and prediction,

    J. Zhang, S. Zheng, W. Chen, G. Du, Q. Fu, and H. Jiang, “A scheme combining feature fusion and hybrid deep learning models for epileptic seizure detection and prediction,” Scientific Reports, vol. 14, p. 16916, July 2024

    2024

  4. [12]

    Residual and bidirectional LSTM for epileptic seizure detection,

    W. Zhao, W.-F. Wang, L. M. Patnaik, B.-C. Zhang, S.-J. Weng, S.-X. Xiao, D.-Z. Wei, and H.-F. Zhou, “Residual and bidirectional LSTM for epileptic seizure detection,” Fron- tiers in Computational Neuroscience , vol. 18, p. 1415967, June 2024

    2024

  5. [13]

    The necessity of Leave One Subject Out 30 (LOSO) cross validation for EEG disease diagnosis,

    S. Kunjan, T. S. Grummett, K. J. Pope, D. M. W. Powers, S. P. Fitzgibbon, T. Bas- tiampillai, M. Battersby, and T. W. Lewis, “The necessity of Leave One Subject Out 30 (LOSO) cross validation for EEG disease diagnosis,” in Brain Informatics (M. Mahmud, M. S. Kaiser, S. Vassane...

    2021

  6. [14]

    Performance evaluation of classification algorithms by k-fold and leave- one-out cross validation,

    T.-T. Wong, “Performance evaluation of classification algorithms by k-fold and leave- one-out cross validation,” Pattern Recognition, vol. 48, no. 9, pp. 2839–2846, 2015

    2015

  7. [15]

    In- dications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state,

    R. G. Andrzejak, K. Lehnertz, F. Mormann, C. Rieke, P. David, and C. E. Elger, “In- dications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state,” Phys. Rev. E, vol. 64, p. 06...

    2001

  8. [16]

    EEGWaveNet: Multiscale CNN-based spatiotemporal feature extraction for EEG seizure detection,

    P. Thuwajit, P. Rangpong, P. Sawangjai, P. Autthasan, R. Chaisaen, N. Banluesom- batkul, P. Boonchit, N. Tatsaringkansakul, T. Sudhawiyangkul, and T. Wilaiprasit- porn, “EEGWaveNet: Multiscale CNN-based spatiotemporal feature extraction for EEG seizure detection,” IEEE Transac...

    2022

  9. [17]

    A brief summary of EEG artifact handling,

    I. Kaya, “A brief summary of EEG artifact handling,” in Brain-Computer Interface (V. Asadpour, ed.), ch. 2, Rijeka: IntechOpen, 2021

    2021

  10. [18]

    Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals,

    U. R. Acharya, S. L. Oh, Y. Hagiwara, J. H. Tan, and H. Adeli, “Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals,” Computers in Biology and Medicine , vol. 100, pp. 270–278, 9 2018

    2018

  11. [19]

    An automated system for epilepsy detection using EEG brain signals based on deep learning approach,

    I. Ullah, M. Hussain, E. Qazi, and H. Aboalsamh, “An automated system for epilepsy detection using EEG brain signals based on deep learning approach,” Expert Systems with Applications, vol. 107, pp. 61–71, Oct. 2018

    2018

  12. [20]

    Patient-independent seizure detection based on channel- perturbation convolutional neural network and bidirectional long short-term memory,

    G. Liu, L. Tian, and W. Zhou, “Patient-independent seizure detection based on channel- perturbation convolutional neural network and bidirectional long short-term memory,” International Journal of Neural Systems , vol. 32, p. 2150051, June 2022

    2022

  13. [21]

    Unsupervised EEG artifact detection and correction,

    S. Saba-Sadiya, E. Chantland, T. Alhanai, T. Liu, and M. M. Ghassemi, “Unsupervised EEG artifact detection and correction,” Frontiers in Digital Health , vol. 2, 2021

    2021

  14. [22]

    Graph eigen decomposition- based feature-selection method for epileptic seizure detection using electroencephalogra- phy,

    M. K. I. Molla, K. M. Hassan, M. R. Islam, and T. Tanaka, “Graph eigen decomposition- based feature-selection method for epileptic seizure detection using electroencephalogra- phy,” Sensors, vol. 20, no. 16, 2020

    2020

  15. [23]

    Multiband entropy–based feature–extraction method for automatic identification of epileptic focus based on high-frequency components in interictal iEEG,

    M. S. Akter, M. R. Islam, Y. Iimura, H. Sugano, K. Fukumori, D. Wang, T. Tanaka, and A. Cichocki, “Multiband entropy–based feature–extraction method for automatic identification of epileptic focus based on high-frequency components in interictal iEEG,” Scientific reports, vol....

    2020

  16. [24]

    J. M. Stern, Atlas of EEG Patterns . Philadelphia, PA: Lippincott Williams & Wilkins, 2005

    2005

  17. [25]

    Improving the generalization of patient non-specific model for epileptic seizure detection,

    G. M. Sigsgaard and Y. Gu, “Improving the generalization of patient non-specific model for epileptic seizure detection,” Biomedical Physics & Engineering Express , vol. 10, p. 15010, 12 2023

    2023

  18. [26]

    Riemannian approaches in brain-computer interfaces: A review,

    F. Yger, M. Berar, and F. Lotte, “Riemannian approaches in brain-computer interfaces: A review,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 25, pp. 1753–1762, 10 2017. 31

    2017

  19. [27]

    Detect- ing EEG outliers for BCI on the Riemannian manifold using spectral clustering,

    M. S. Yamamoto, K. Sadatnejad, T. Tanaka, R. Islam, Y. Tanaka, and F. Lotte, “Detect- ing EEG outliers for BCI on the Riemannian manifold using spectral clustering,” in 2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pp. ...

    2020

  20. [28]

    Active selection of source patients in transfer learning for epileptic seizure detection using riemannian manifold,

    T. Orihara, K. M. Hassan, and T. Tanaka, “Active selection of source patients in transfer learning for epileptic seizure detection using riemannian manifold,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5, 2023

    2023

  21. [29]

    Detection of epileptic seizures in long EEG recordings using an anomaly detector with artifact rejection,

    K. M. Hassan, X. Zhao, H. Sugano, and T. Tanaka, “Detection of epileptic seizures in long EEG recordings using an anomaly detector with artifact rejection,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2230–2234, 2024

    2024

  22. [30]

    Riemannian manifold-based epilep- tic seizure detection using transfer learning and artifact rejection techniques,

    K. M. Hassan, X. Zhao, H. Sugano, and T. Tanaka, “Riemannian manifold-based epilep- tic seizure detection using transfer learning and artifact rejection techniques,” APSIPA Transactions on Signal and Information Processing , vol. 13, no. 1, 2024

    2024

  23. [31]

    Seizure detection by convolutional neural network-based analysis of scalp electroencephalography plot images,

    A. Emami, N. Kunii, T. Matsuo, T. Shinozaki, K. Kawai, and H. Takahashi, “Seizure detection by convolutional neural network-based analysis of scalp electroencephalography plot images,” NeuroImage: Clinical , vol. 22, p. 101684, 2019

    2019

  24. [32]

    Convolu- tional neural network with autoencoder-assisted multiclass labelling for seizure detection based on scalp electroencephalography,

    H. Takahashi, A. Emami, T. Shinozaki, N. Kunii, T. Matsuo, and K. Kawai, “Convolu- tional neural network with autoencoder-assisted multiclass labelling for seizure detection based on scalp electroencephalography,” Computers in Biology and Medicine , vol. 125, p. 104016, 2020

    2020

  25. [33]

    A real-time epilepsy seizure detection approach based on EEG using short-time Fourier transform and Google-Net convolutional neural network,

    M. Shen, F. Yang, P. Wen, B. Song, and Y. Li, “A real-time epilepsy seizure detection approach based on EEG using short-time Fourier transform and Google-Net convolutional neural network,” Heliyon, vol. 10, no. 11, p. e31827, 2024

    2024

  26. [34]

    Epileptic seizure detection from electroencephalo- gram (EEG) signals using linear graph convolutional network and DenseNet based hy- brid framework,

    F. A. Jibon, M. H. Miraz, M. U. Khandaker, M. Rashdan, M. Salman, A. Tasbir, N. H. Nishar, and F. H. Siddiqui, “Epileptic seizure detection from electroencephalo- gram (EEG) signals using linear graph convolutional network and DenseNet based hy- brid framework,” Journal of Rad...

    2023

  27. [35]

    EEGNet: A compact convolutional neural network for EEG-based brain-computer in- terfaces,

    V. J. Lawhern, N. Solon, N. R. Waytowich, S. M. Gordon, C. P. Hung, and B. J. Lance, “EEGNet: A compact convolutional neural network for EEG-based brain-computer in- terfaces,” Journal of neural engineering , vol. 15, p. 056013, July 2018

    2018

  28. [36]

    EEG Conformer: Convolutional transformer for EEG decoding and visualization,

    Y. Song, Q. Zheng, B. Liu, and X. Gao, “EEG Conformer: Convolutional transformer for EEG decoding and visualization,” IEEE Transactions on Neural Systems and Reha- bilitation Engineering, vol. 31, pp. 710–719, 2023

    2023

  29. [37]

    BIOT: Biosignal transformer for cross-data learning in the wild,

    C. Yang, M. Westover, and J. Sun, “BIOT: Biosignal transformer for cross-data learning in the wild,” in Advances in Neural Information Processing Systems(A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, eds.), vol. 36, pp. 78240–78260, Curran Associates, Inc., 2023

    2023

  30. [38]

    Large brain model for learning generic repre- sentations with tremendous EEG data in BCI,

    W.-B. Jiang, L.-M. Zhao, and B.-L. Lu, “Large brain model for learning generic repre- sentations with tremendous EEG data in BCI,” in The Twelfth International Conference on Learning Representations, 2024. 32

    2024

  31. [39]

    CBraMod: A criss-cross brain foundation model for EEG decoding,

    J. Wang, S. Zhao, Z. Luo, Y. Zhou, H. Jiang, S. Li, T. Li, and G. Pan, “CBraMod: A criss-cross brain foundation model for EEG decoding,” in The Thirteenth International Conference on Learning Representations, 2025

    2025

  32. [40]

    Low latency real-time seizure detection using transfer deep learning,

    V. Khalkhali, N. Shawki, V. Shah, M. Golmohammadi, I. Obeid, and J. Picone, “Low latency real-time seizure detection using transfer deep learning,” in 2021 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), (Philadelphia, PA, USA), pp. 1– 7, IEEE, Dec. 2021

    2021

  33. [41]

    Epileptic seizure detection in clinical EEGs using an XGboost- based method,

    L. Wei and C. Mooney, “Epileptic seizure detection in clinical EEGs using an XGboost- based method,” in 2020 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), pp. 1–6, 2020

    2020

  34. [42]

    Duration of epileptic seizure types: A data-driven approach,

    P. Meritam Larsen, S. W¨ ustenhagen, D. Terney, E. Gardella, H. Aurlien, and S. Beniczky, “Duration of epileptic seizure types: A data-driven approach,” Epilepsia, vol. 64, pp. 469– 478, Feb. 2023

    2023

  35. [43]

    The Temple University Hospital EEG data corpus,

    I. Obeid and J. Picone, “The Temple University Hospital EEG data corpus,” Frontiers in Neuroscience, vol. Volume 10 - 2016, 2016

    2016

  36. [44]

    EEG synchronization analysis for seizure prediction: A study on data of noninvasive recordings,

    P. Detti, G. Vatti, and G. Zabalo Manrique De Lara, “EEG synchronization analysis for seizure prediction: A study on data of noninvasive recordings,” Processes, vol. 8, p. 846, July 2020

    2020

  37. [45]

    Siena Scalp EEG Database

    P. Detti, “Siena Scalp EEG Database.” https://doi.org/10.13026/5d4a-j060

    doi

  38. [46]

    ECOD: Unsupervised outlier detection using empirical cumulative distribution functions,

    Z. Li, Y. Zhao, X. Hu, N. Botta, C. Ionescu, and G. Chen, “ECOD: Unsupervised outlier detection using empirical cumulative distribution functions,” IEEE Transactions on Knowledge and Data Engineering , pp. 1–1, 2022

    2022

  39. [47]

    Pyod 2: A python library for outlier detection with LLM-powered model selection,

    S. Chen, Z. Qian, W. Siu, X. Hu, J. Li, S. Li, Y. Qin, T. Yang, Z. Xiao, W. Ye, Y. Zhang, Y. Dong, and Y. Zhao, “Pyod 2: A python library for outlier detection with LLM-powered model selection,” arXiv preprint arXiv:2412.12154 , 2024

    2024 arXiv

  40. [48]

    PyOD: A python toolbox for scalable outlier detec- tion,

    Y. Zhao, Z. Nasrullah, and Z. Li, “PyOD: A python toolbox for scalable outlier detec- tion,” Journal of Machine Learning Research , vol. 20, no. 96, pp. 1–7, 2019

    2019

  41. [49]

    J. L. Fern´ andez-Torre,Interictal EEG , pp. 701–712. London: Springer London, 2010

    2010

  42. [50]

    LightGBM: a highly efficient gradient boosting decision tree,

    G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T. Liu, “LightGBM: a highly efficient gradient boosting decision tree,” in Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, (Red Hook, NY, USA), p. 3149–3157, Curr...

    2017

  43. [51]

    Methods for artifact detection and removal from scalp EEG: A review,

    M. K. Islam, A. Rastegarnia, and Z. Yang, “Methods for artifact detection and removal from scalp EEG: A review,” Neurophysiologie Clinique/Clinical Neurophysiology, vol. 46, pp. 287–305, Nov. 2016

    2016

  44. [52]

    American Clinical Neurophysiology Society Guideline 1: Minimum Technical Requirements for Performing Clinical Elec- troencephalography,

    S. R. Sinha, L. Sullivan, D. Sabau, D. San-Juan, K. E. Dombrowski, J. J. Halford, A. J. Hani, F. W. Drislane, and M. M. Stecker, “American Clinical Neurophysiology Society Guideline 1: Minimum Technical Requirements for Performing Clinical Elec- troencephalography,” Journal of...

    2016

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.