pith. sign in

arxiv: 2309.03147 · v4 · pith:KRQJIBIRnew · submitted 2023-09-06 · 📡 eess.SP

Real-Time Non-Invasive Imaging and Detection of Spreading Depolarizations through EEG: An Ultra-Light Explainable Deep Learning Approach

Yinzhe Wu , Sharon Jewell , Xiaodan Xing , Yang Nan , Anthony J. Strong , Guang Yang , Martyn G. Boutelle This is my paper

Pith reviewed 2026-05-24 06:51 UTC · model grok-4.3

classification 📡 eess.SP
keywords spreading depolarizationsEEGspectrogramdeep learningnon-invasive detectionmulti-modal networkreal-time processingbrain injury
0
0 comments X

The pith

A lightweight multi-modal network fuses EEG spectrograms with temporal power vectors to detect spreading depolarizations non-invasively and in real time.

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

The paper sets out to demonstrate that recasting spreading depolarization detection as a sequential 2-D imaging task on EEG spectrograms, then fusing those images with temporal power vectors inside an ultra-light network, raises identification accuracy above what any single electrode can achieve. This design removes the need for fixed high-density electrode arrays and cuts processing time from two hours to under 0.3 seconds. A sympathetic reader would care because spreading depolarizations are a treatable driver of secondary brain injury, so faster, non-invasive, and flexible monitoring could shift neurocritical care from delayed diagnosis to immediate intervention. The authors treat the added frequency dimension visible on spectrograms as the source of the performance gain.

Core claim

The central claim is that transforming the SD identification problem from a 1-D time-series task into a sequential 2-D rendered imaging task, then feeding both the spectrogram images and temporal power vectors into an ultra-light multi-modal network, improves accuracy over single-electrode analysis, supports flexible and low-density electrode maps, delivers inference in less than 0.3 seconds, and supplies preliminary evidence that frequency profiles contain useful discriminative information for SDs.

What carries the argument

The ultra-light multi-modal network that fuses EEG spectrogram imaging with temporal power vectors to process sequential 2-D rendered data.

If this is right

  • Detection becomes feasible on ultra-low-density EEG setups with variable electrode positions.
  • Inference completes in under 0.3 seconds, replacing conventional two-hour processing.
  • Accuracy rises above single-electrode performance through the combined modalities.
  • The frequency dimension on spectrograms can be exploited to improve SD identification.

Where Pith is reading between the lines

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

  • Portable low-channel EEG devices could support continuous bedside monitoring in intensive-care settings.
  • The same spectrogram-plus-power fusion might apply to other scalp-detectable neurological events that carry frequency signatures.
  • Large-scale validation across patients with differing noise profiles would be required before clinical deployment.

Load-bearing premise

The frequency content shown on EEG spectrograms supplies reliable discriminative features for spreading depolarizations that remain detectable despite the noise and attenuation of scalp recordings.

What would settle it

A controlled test in which a model using only temporal power vectors matches or exceeds the accuracy of the full multi-modal version that adds spectrogram images would falsify the claimed value of the frequency dimension.

Figures

Figures reproduced from arXiv: 2309.03147 by Anthony J. Strong, Guang Yang, Martyn G. Boutelle, Sharon Jewell, Xiaodan Xing, Yang Nan, Yinzhe Wu.

Figure 1
Figure 1. Figure 1: Challenges and our solutions to non-invasive detection of spreading depolarization (SD). (A) SD-specific features in the AC band tracing and power of EEG. (B) Challenges in defining SDs in EEG and limitations of the current methods [30], [33]: The conventional method which determines SD based on reduction of power temporally is prone to often fluctuating EEG AC power offsets in real-world patients, produci… view at source ↗
Figure 2
Figure 2. Figure 2: Examples of a SD occurrence in (1) near-direct current (DC)/DC (0.0001-0.01Hz) and (2) alternating current (AC) (0.5-45Hz) bands of (a) an EEG tracing and (b) an ECoG tracing for an SD occurrence. Axes: Time (Horizontal) and Power/mV (Vertical). A SD can be clearly identified from ECoG DC as a large negative slow potential change (SPC) and from ECoG AC as a drastic amplitude and power reduction. However, f… view at source ↗
Figure 3
Figure 3. Figure 3: The general framework for identification of SD over an EEG tracing from a single electrode. (Stage 1) The EEG of a scalp electrode first (A) undergoes bandpass to obtain (Stage 2) its AC segment and then calculated for (Stage 3) its (3.1) spectrogram images and (3.2) signal power vectors by (B.1) short-time Fourier transform and (B.2) overall power calculations in AC band respectively. These spectrogram im… view at source ↗
Figure 5
Figure 5. Figure 5: A binary outcome at Stage 5 is produced after the CNN fusion layer ( [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
read the original abstract

A core aim of neurocritical care is to prevent secondary brain injury. Spreading depolarizations (SDs) have been identified as an important independent cause of secondary brain injury. SDs are usually detected using invasive electrocorticography recorded at high sampling frequency. Recent pilot studies suggest a possible utility of scalp electrodes generated electroencephalogram (EEG) for non-invasive SD detection. However, noise and attenuation of EEG signals makes this detection task extremely challenging. Previous methods focus on detecting temporal power change of EEG over a fixed high-density map of scalp electrodes, which is not always clinically feasible. Having a specialized spectrogram as an input to the automatic SD detection model, this study is the first to transform SD identification problem from a detection task on a 1-D time-series wave to a task on a sequential 2-D rendered imaging. This study presented a novel ultra-light-weight multi-modal deep-learning network to fuse EEG spectrogram imaging and temporal power vectors to enhance SD identification accuracy over each single electrode, allowing flexible EEG map and paving the way for SD detection on ultra-low-density EEG with variable electrode positioning. Our proposed model has an ultra-fast processing speed (<0.3 sec). Compared to the conventional methods (2 hours), this is a huge advancement towards early SD detection and to facilitate instant brain injury prognosis. Seeing SDs with a new dimension - frequency on spectrograms, we demonstrated that such additional dimension could improve SD detection accuracy, providing preliminary evidence to support the hypothesis that SDs may show implicit features over the frequency profile.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper claims to introduce the first transformation of SD detection from 1D EEG time-series to sequential 2D spectrogram imaging, presenting an ultra-light multi-modal DL network that fuses spectrogram images with temporal power vectors. This is asserted to improve accuracy over single electrodes, support flexible low-density EEG maps, achieve <0.3s inference (vs. conventional 2-hour methods), and demonstrate that the frequency dimension on spectrograms supplies useful discriminative features for SDs despite scalp EEG noise and attenuation.

Significance. If the accuracy gains and ablation evidence hold under proper validation, the work would be significant for enabling real-time non-invasive SD monitoring in neurocritical care, with the ultra-light model and sub-second speed offering clear clinical advantages over invasive ECoG or slow conventional processing. The framing as an empirical modeling contribution with potential for variable electrode positioning is a strength.

major comments (2)
  1. Abstract: the central claim that 'such additional dimension could improve SD detection accuracy' and that spectrogram features remain discriminative after noise/attenuation is unsupported by any quantitative metrics, dataset sizes, cross-validation details, or ablation results isolating the spectrogram branch versus temporal vectors alone. This renders the accuracy improvement unevaluable.
  2. Abstract and methods framing: no details are supplied on model architecture, training procedure, loss functions, or validation strategy for the ultra-light multi-modal network, which is load-bearing for assessing both the claimed speed (<0.3 sec) and the assertion of improvement over single-electrode baselines.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address each major comment below, indicating where the manuscript already provides supporting information and where we will make revisions for clarity and completeness.

read point-by-point responses

  1. Referee: [—] Abstract: the central claim that 'such additional dimension could improve SD detection accuracy' and that spectrogram features remain discriminative after noise/attenuation is unsupported by any quantitative metrics, dataset sizes, cross-validation details, or ablation results isolating the spectrogram branch versus temporal vectors alone. This renders the accuracy improvement unevaluable.

    Authors: The abstract is a concise summary and does not repeat the full quantitative results. The manuscript reports dataset sizes, cross-validation procedures, and ablation studies isolating the spectrogram branch contribution in the Results section, along with accuracy metrics showing improvement over single-electrode and single-modality baselines. We will revise the abstract to include key quantitative figures (e.g., accuracy gains, inference latency, and dataset scale) and a brief reference to the ablation evidence so the central claims are directly supported and evaluable from the abstract. revision: yes

  2. Referee: [—] Abstract and methods framing: no details are supplied on model architecture, training procedure, loss functions, or validation strategy for the ultra-light multi-modal network, which is load-bearing for assessing both the claimed speed (<0.3 sec) and the assertion of improvement over single-electrode baselines.

    Authors: The Methods section describes the ultra-light multi-modal architecture (spectrogram CNN branch fused with temporal power vector branch), training procedure, loss function, and validation strategy, with the reported sub-0.3 s inference tied directly to the model design. To address the concern about framing, we will add a concise summary paragraph and a hyperparameter table in the Methods section and update the abstract to explicitly link the speed claim to the architecture details. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical modeling contribution is self-contained

full rationale

The paper describes an empirical deep-learning architecture for SD detection that fuses spectrogram images with temporal power vectors. No equations, fitted parameters, or derivation chain are presented that would reduce reported accuracies or claims to inputs by construction. The work makes no load-bearing appeals to self-citations, uniqueness theorems, or ansatzes imported from prior author work; performance claims rest on standard supervised training and evaluation rather than self-referential definitions. The absence of any mathematical reduction or circular step keeps the circularity score at the default non-finding level.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim rests on the empirical premise that spectrogram frequency content supplies additional discriminative power for SDs in scalp EEG. No free parameters, axioms, or invented entities are stated in the abstract.

pith-pipeline@v0.9.0 · 5839 in / 1273 out tokens · 14951 ms · 2026-05-24T06:51:46.053180+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

65 extracted references · 65 canonical work pages

  1. [1]

    Spreading and Synchronous Depressions of Cortical Activity in Acutely Injured Human Brain,

    A. J. Strong et al., “Spreading and Synchronous Depressions of Cortical Activity in Acutely Injured Human Brain,” Stroke, vol. 33, no. 12, pp. 2738–2743, Dec. 2002, doi: 10.1161/01.STR.0000043073.69602.09

    work page doi:10.1161/01.str.0000043073.69602.09 2002

  2. [2]

    SPREADING DEPRESSION OF ACTIVITY IN THE CEREBRAL CORTEX,

    A. A. P. Leao, “SPREADING DEPRESSION OF ACTIVITY IN THE CEREBRAL CORTEX,” J Neurophysiol, vol. 7, no. 6, pp. 359– 390, Nov. 1944, doi: 10.1152/jn.1944.7.6.359

    work page doi:10.1152/jn.1944.7.6.359 1944

  3. [3]

    Spreading Depolarizations and Late Secondary Insults after Traumatic Brain Injury,

    J. A. Hartings et al., “Spreading Depolarizations and Late Secondary Insults after Traumatic Brain Injury,” J Neurotrauma, vol. 26, no. 11, pp. 1857–1866, Nov. 2009, doi: 10.1089/neu.2009.0961

    work page doi:10.1089/neu.2009.0961 2009

  4. [4]

    Cortical spreading depression and peri- infarct depolarization in acutely injured human cerebral cortex,

    M. Fabricius et al. , “Cortical spreading depression and peri- infarct depolarization in acutely injured human cerebral cortex,” Brain, vol. 129, no. 3, pp. 778–790, Mar. 2006, doi: 10.1093/brain/awh716

    work page doi:10.1093/brain/awh716 2006

  5. [5]

    Delayed ischaemic neurological deficits after subarachnoid haemorrhage are associated with clusters of spreading depolarizations,

    J. P. Dreier et al. , “Delayed ischaemic neurological deficits after subarachnoid haemorrhage are associated with clusters of spreading depolarizations,” Brain, vol. 129, no. 12, pp. 3224 –3237, Jun. 2006, doi: 10.1093/brain/awl297

    work page doi:10.1093/brain/awl297 2006

  6. [6]

    Spreading depolarizations occur in human ischemic stroke with high incidence,

    C. Dohmen et al. , “Spreading depolarizations occur in human ischemic stroke with high incidence,” Ann Neurol, vol. 63, no. 6, pp. 720–728, Jun. 2008, doi: 10.1002/ana.21390

    work page doi:10.1002/ana.21390 2008

  7. [7]

    Spreading depolarizations in patients with spontaneous intracerebral hemorrhage: Association with perihematomal edema progression,

    R. Helbok et al. , “Spreading depolarizations in patients with spontaneous intracerebral hemorrhage: Association with perihematomal edema progression,” Journal of Cerebral Blood Flow & Metabolism, vol. 37, no. 5, pp. 1871 –1882, May 2017, doi: 10.1177/0271678X16651269

    work page doi:10.1177/0271678x16651269 2017

  8. [8]

    Clinical Relevance of Cortical Spreading Depression in Neurological Disorders: Migraine, Malignant Stroke, Subarachnoid and Intracranial Hemorrhage, and Traumatic Brain Injury,

    M. Lauritzen, J. P. Dreier, M. Fabricius, J. A. Hartings, R. Graf, and A. J. Strong, “Clinical Relevance of Cortical Spreading Depression in Neurological Disorders: Migraine, Malignant Stroke, Subarachnoid and Intracranial Hemorrhage, and Traumatic Brain Injury,” Journal of Cerebral Blood Flow & Metabolism , vol. 31, no. 1, pp. 17–35, Jan. 2011, doi: 10.1...

    work page doi:10.1038/jcbfm.2010.191 2011

  9. [9]

    Propagation of cortical spreading depolarization in the human cortex after malignant stroke,

    J. Woitzik et al., “Propagation of cortical spreading depolarization in the human cortex after malignant stroke,” Neurology, vol. 80, no. 12, pp. 1095–1102, Mar. 2013, doi: 10.1212/WNL.0b013e3182886932

    work page doi:10.1212/wnl.0b013e3182886932 2013

  10. [10]

    Mechanisms of Spreading Depression and Hypoxic Spreading Depression-Like Depolarization,

    G. G. Somjen, “Mechanisms of Spreading Depression and Hypoxic Spreading Depression-Like Depolarization,” Physiol Rev, vol. 81, no. 3, pp. 1065–1096, Jul. 2001, doi: 10.1152/physrev.2001.81.3.1065

    work page doi:10.1152/physrev.2001.81.3.1065 2001

  11. [11]

    The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease,

    J. P. Dreier, “The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease,” Nature Medicine, vol. 17, no. 4. pp. 439– 447, Apr. 2011. doi: 10.1038/nm.2333

    work page doi:10.1038/nm.2333 2011

  12. [12]

    High -Resolution In Vivo Imaging of the Neurovascular Unit during Spreading Depression,

    J. Chuquet, L. Hollender, and E. A. Nimchinsky, “High -Resolution In Vivo Imaging of the Neurovascular Unit during Spreading Depression,” Journal of Neuroscience, vol. 27, no. 15, pp. 4036–4044, Apr. 2007, doi: 10.1523/JNEUROSCI.0721-07.2007

    work page doi:10.1523/jneurosci.0721-07.2007 2007

  13. [13]

    Different Mechanisms Promote Astrocyte Ca 2+ Waves and Spreading Depression in the Mouse Neocortex,

    O. Peters, C. G. Schipke, Y. Hashimoto, and H. Kettenmann, “Different Mechanisms Promote Astrocyte Ca 2+ Waves and Spreading Depression in the Mouse Neocortex,” The Journal of Neuroscience, vol. 23, no. 30, pp. 9888– 9896, Oct. 2003, doi: 10.1523/JNEUROSCI.23-30-09888.2003

    work page doi:10.1523/jneurosci.23-30-09888.2003 2003

  14. [14]

    Ion changes in spreading ischaemia induce rat middle cerebral artery constriction in the absence of NO,

    O. Windmüller et al., “Ion changes in spreading ischaemia induce rat middle cerebral artery constriction in the absence of NO,” Brain, vol. 128, no. 9, pp. 2042–2051, Sep. 2005, doi: 10.1093/brain/awh545

    work page doi:10.1093/brain/awh545 2042

  15. [15]

    ‘Spreading depression of Leão’ and its emerging relevance to acute brain injury in humans,

    M. Lauritzen and A. J. Strong, “‘Spreading depression of Leão’ and its emerging relevance to acute brain injury in humans,” Journal of Cerebral Blood Flow & Metabolism , vol. 37, no. 5, pp. 1553 –1570, May 2017, doi: 10.1177/0271678X16657092

    work page doi:10.1177/0271678x16657092 2017

  16. [16]

    The continuum of spreading depolarizations in acute cortical lesion development: Examining Leão’s legacy,

    J. A. Hartings et al., “The continuum of spreading depolarizations in acute cortical lesion development: Examining Leão’s legacy,” Journal of Cerebral Blood Flow and Metabolism , vol. 37, no. 5. Nature Publishing Group, pp. 1571– 1594, May 01, 2017. doi: 10.1177/0271678X16654495

    work page doi:10.1177/0271678x16654495 2017

  17. [17]

    Prognostic Value of Spreading Depolarizations in Patients With Severe Traumatic Brain Injury,

    J. A. Hartings et al., “Prognostic Value of Spreading Depolarizations in Patients With Severe Traumatic Brain Injury,” JAMA Neurol, vol. 77, no. 4, p. 489, Apr. 2020, doi: 10.1001/jamaneurol.2019.4476

    work page doi:10.1001/jamaneurol.2019.4476 2020

  18. [18]

    Spreading depolarisations and outcome after traumatic brain injury: a prospective observational study,

    J. A. Hartings et al. , “Spreading depolarisations and outcome after traumatic brain injury: a prospective observational study,” Lancet Neurol, vol. 10, no. 12, pp. 1058– 1064, Dec. 2011, doi: 10.1016/S1474-4422(11)70243-5

    work page doi:10.1016/s1474-4422(11)70243-5 2011

  19. [19]

    The Stroke- Migraine Depolarization Continuum,

    J. P. Dreier and C. Reiffurth , “The Stroke- Migraine Depolarization Continuum,” Neuron, vol. 86, no. 4, pp. 902– 922, May 2015, doi: 10.1016/j.neuron.2015.04.004

    work page doi:10.1016/j.neuron.2015.04.004 2015

  20. [20]

    Recording, analysis, and interpretation of spreading depolarizations in neurointensive care: Review and recommendations of the COSBID research group,

    J. P. Dreier et al. , “Recording, analysis, and interpretation of spreading depolarizations in neurointensive care: Review and recommendations of the COSBID research group,” Journal of Cerebral Blood Flow and Metabolism , vol. 37, no. 5. Nature Publishing Group, pp. 1595– 1625, May 01, 2017. doi: 10.1177/0271678X16654496

    work page doi:10.1177/0271678x16654496 2017

  21. [21]

    SPREADING DEPRESSION, SPREADING DEPOLARIZATIONS, AND THE CEREBRAL VASCULATURE,

    C. Ayata and M. Lauritzen, “SPREADING DEPRESSION, SPREADING DEPOLARIZATIONS, AND THE CEREBRAL VASCULATURE,” Physiol Rev, vol. 95, pp. 953– 993, 2015, doi: 10.1152/physrev.00027.2014.-Spreading

    work page doi:10.1152/physrev.00027.2014.-spreading 2015

  22. [22]

    Detection of spreading depolarization with intraparenchymal electrodes in the injured human brain,

    T. Jeffcote et al. , “Detection of spreading depolarization with intraparenchymal electrodes in the injured human brain,” Neurocrit Care, vol. 20, no. 1, pp. 21–31, 2014, doi: 10.1007/s12028-013-9938- 7

    work page doi:10.1007/s12028-013-9938- 2014

  23. [23]

    Development and Evaluation of a Method for Automated Detection of Spreading Depolarizations in the Injured Human Brain,

    S. Jewell et al. , “Development and Evaluation of a Method for Automated Detection of Spreading Depolarizations in the Injured Human Brain,” Neurocrit Care, vol. 35, pp. 160–175, Oct. 2021, doi: 10.1007/s12028-021-01228-x

    work page doi:10.1007/s12028-021-01228-x 2021

  24. [24]

    Correlates of spreading depolarization in human scalp electroencephalography,

    C. Drenckhahn et al. , “Correlates of spreading depolarization in human scalp electroencephalography,” Brain , vol. 135, no. 3, pp. 853–868, 2012, doi: 10.1093/brain/aws010

    work page doi:10.1093/brain/aws010 2012

  25. [25]

    Spreading depression in continuous electroencephalography of brain trauma,

    J. A. Hartings et al. , “Spreading depression in continuous electroencephalography of brain trauma,” Ann Neurol, vol. 76, no. 5, pp. 681–694, Nov. 2014, doi: 10.1002/ana.24256

    work page doi:10.1002/ana.24256 2014

  26. [26]

    Cortical Spreading Depolarizations and Clinically Measured Scalp EEG Activity After Aneurysmal Subarachnoid Hemorrhage and Traumatic Brain Injury,

    S. Sivakumar et al. , “Cortical Spreading Depolarizations and Clinically Measured Scalp EEG Activity After Aneurysmal Subarachnoid Hemorrhage and Traumatic Brain Injury,” Neurocrit Care, vol. 37, no. S1, pp. 49– 59, Jun. 2022, doi: 10.1007/s12028- 021-01418-7

    work page doi:10.1007/s12028- 2022

  27. [27]

    Commentary: Detecting Cortical Spreading Depolarization with Full Band Scalp Electroencephalography: An Illusion?,

    J. A. Hartings, L. B. Ngwenya, T. Watanabe, and B. Foreman, “Commentary: Detecting Cortical Spreading Depolarization with Full Band Scalp Electroencephalography: An Illusion?,” Front Syst Neurosci, vol. 12, May 2018, doi: 10.3389/fnsys.2018.00019

    work page doi:10.3389/fnsys.2018.00019 2018

  28. [28]

    Detecting Cortical Spreading Depolarization with Full Band Scalp Electroencephalography: An Illusion?,

    J. Hofmeijer, C. R. van Kaam, B. van de Werff, S. E. Vermeer, M. C. Tjepkema-Cloostermans, and M. J. A. M. van Putten, “Detecting Cortical Spreading Depolarization with Full Band Scalp Electroencephalography: An Illusion?,” Front Neurol , vol. 9, Jan. 2018, doi: 10.3389/fneur.2018.00017

    work page doi:10.3389/fneur.2018.00017 2018

  29. [29]

    First Report of Spreading Depolarization Correlates on Scalp EEG Confirmed with a Depth Electrode,

    D. Robinson, J. Hartings, and B. Foreman, “First Report of Spreading Depolarization Correlates on Scalp EEG Confirmed with a Depth Electrode,” Neurocrit Care, vol. 35, no. S2, pp. 100–104, Oct. 2021, doi: 10.1007/s12028-021-01360-8. 12

    work page doi:10.1007/s12028-021-01360-8 2021

  30. [30]

    An Algorithm for Automated, Noninvasive Detection of Cortical Spreading Depolarizations Based on EEG Simulations,

    A. Chamanzar et al. , “An Algorithm for Automated, Noninvasive Detection of Cortical Spreading Depolarizations Based on EEG Simulations,” IEEE Trans Biomed Eng, vol. 66, no. 4, pp. 1115–1126, Apr. 2019, doi: 10.1109/TBME.2018.2867112

    work page doi:10.1109/tbme.2018.2867112 2019

  31. [31]

    Non -invasive monitoring of spreading depression,

    Z. J. R. Bastany, S. Askari, G. A. Dumont, E. -J. Speckmann, and A. Gorji, “Non -invasive monitoring of spreading depression,” Neuroscience, vol. 333, pp. 1– 12, Oct. 2016, doi: 10.1016/j.neuroscience.2016.06.056

    work page doi:10.1016/j.neuroscience.2016.06.056 2016

  32. [32]

    Development and implementation of real- time excitability measurement in the injured human brain: a novel neurocritical care monitoring tool.,

    S. L. Jewell, “Development and implementation of real- time excitability measurement in the injured human brain: a novel neurocritical care monitoring tool.,” Imperial College London, London, 2021

    work page 2021

  33. [33]

    Noninvasive and reliable automated detection of spreading depolarization in severe traumatic brain injury using scalp EEG,

    A. Chamanzar, J. Elmer, L. Shutter, J. Hartings, and P. Grover, “Noninvasive and reliable automated detection of spreading depolarization in severe traumatic brain injury using scalp EEG,” Communications Medicine, vol. 3, no. 1, p. 113, Aug. 2023, doi: 10.1038/s43856-023-00344-3

    work page doi:10.1038/s43856-023-00344-3 2023

  34. [34]

    Automated, Scalable and Generalizable Deep Learning for Tracking Cortical Spreading Depression Using EEG,

    A. Chamanzar, X. Liu, L. Y. Jiang, K. A. Vogt, J. M. F. Moura, and P. Grover, “Automated, Scalable and Generalizable Deep Learning for Tracking Cortical Spreading Depression Using EEG,” in 2021 10th International IEEE/EMBS Conference on Neural Engineering (NER), IEEE, May 2021, pp. 416– 419. doi: 10.1109/NER49283.2021.9441333

    work page doi:10.1109/ner49283.2021.9441333 2021

  35. [35]

    Clinical utility of EEG in diagnosing and monitoring epilepsy in adults,

    W. O. Tatum et al. , “Clinical utility of EEG in diagnosing and monitoring epilepsy in adults,” Clinical Neurophysiology, vol. 129, no. 5, pp. 1056–1082, May 2018, doi: 10.1016/j.clinph.2018.01.019

    work page doi:10.1016/j.clinph.2018.01.019 2018

  36. [36]

    Education Research: Neurology Resident EEG Education,

    F. A. Nascimento and J. R. Gavvala, “Education Research: Neurology Resident EEG Education,” Neurology, vol. 96, no. 17, pp. 821–824, Apr. 2021, doi: 10.1212/WNL.0000000000011354

    work page doi:10.1212/wnl.0000000000011354 2021

  37. [37]

    Visual EEG reviewing times with SCORE EEG,

    J. Brogger, T. Eichele, E. Aanestad, H. Olberg, I. Hjelland, and H. Aurlien, “Visual EEG reviewing times with SCORE EEG,” Clin Neurophysiol Pract , vol. 3, pp. 59– 64, 2018, doi: 10.1016/j.cnp.2018.03.002

    work page doi:10.1016/j.cnp.2018.03.002 2018

  38. [38]

    The state of everyday quantitative EEG use in Canada: A national technologist survey,

    M. C. Ng and K. Gillis, “The state of everyday quantitative EEG use in Canada: A national technologist survey,” Seizure, vol. 49, pp. 5–7, Jul. 2017, doi: 10.1016/j.seizure.2017.05.003

    work page doi:10.1016/j.seizure.2017.05.003 2017

  39. [39]

    AI in health and medicine,

    P. Rajpurkar, E. Chen, O. Banerjee, and E. J. Topol, “AI in health and medicine,” Nat Med , vol. 28, no. 1, pp. 31– 38, Jan. 2022, doi: 10.1038/s41591-021-01614-0

    work page doi:10.1038/s41591-021-01614-0 2022

  40. [40]

    Machine learning and wearable devices of the future,

    S. Beniczky, P. Karoly, E. Nurse, P. Ryvlin, and M. Cook, “Machine learning and wearable devices of the future,” Epilepsia, vol. 62, no. S2, Mar. 2021, doi: 10.1111/epi.16555

    work page doi:10.1111/epi.16555 2021

  41. [41]

    Machine learning applications in epilepsy,

    B. Abbasi and D. M. Goldenholz, “Machine learning applications in epilepsy,” Epilepsia, vol. 60, no. 10, pp. 2037– 2047, Oct. 2019, doi: 10.1111/epi.16333

    work page doi:10.1111/epi.16333 2037

  42. [42]

    Seizure detection using EEG: A survey of different techniques,

    M. Z. Ahmad, M. Saeed, S. Saleem, and A. M. Kamboh, “Seizure detection using EEG: A survey of different techniques,” in 2016 International Conference on Emerging Technologies (ICET) , IEEE, Oct. 2016, pp. 1–6. doi: 10.1109/ICET.2016.7813209

    work page doi:10.1109/icet.2016.7813209 2016

  43. [43]

    A Survey and Tutorial of EEG-Based Brain Monitoring for Driver State Analysis,

    C. Zhang and A. Eskandarian, “A Survey and Tutorial of EEG-Based Brain Monitoring for Driver State Analysis,” IEEE/CAA Journal of Automatica Sinica , vol. 8, no. 7, pp. 1222 –1242, Jul. 2021, doi: 10.1109/JAS.2020.1003450

    work page doi:10.1109/jas.2020.1003450 2021

  44. [44]

    Detecting abnormal electroencephalograms using deep convolutional networks,

    K. G. van Leeuwen, H. Sun, M. Tabaeizadeh, A. F. Struck, M. J. A. M. van Putten, and M. B. Westover, “Detecting abnormal electroencephalograms using deep convolutional networks,” Clinical Neurophysiology, vol. 130, no. 1, pp. 77– 84, Jan. 2019, doi: 10.1016/j.clinph.2018.10.012

    work page doi:10.1016/j.clinph.2018.10.012 2019

  45. [45]

    Seizure detection using scalp‐ <scp>EEG</scp>,

    C. Baumgartner and J. P. Koren, “Seizure detection using scalp‐ <scp>EEG</scp>,” Epilepsia, vol. 59, no. S1, pp. 14– 22, Jun. 2018, doi: 10.1111/epi.14052

    work page doi:10.1111/epi.14052 2018

  46. [46]

    A machine- learning algorithm for neonatal seizure recognition: a multicentre, randomised, controlled trial,

    A. M. Pavel et al. , “A machine- learning algorithm for neonatal seizure recognition: a multicentre, randomised, controlled trial,” Lancet Child Adolesc Health, vol. 4, no. 10, pp. 740–749, Oct. 2020, doi: 10.1016/S2352-4642(20)30239-X

    work page doi:10.1016/s2352-4642(20)30239-x 2020

  47. [47]

    Automated detection of absence seizures using a wearable electroencephalographic device: a phase 3 validation study and feasibility of automated behavioral testing,

    G. Japaridze et al., “Automated detection of absence seizures using a wearable electroencephalographic device: a phase 3 validation study and feasibility of automated behavioral testing,” Epilepsia, Mar. 2022, doi: 10.1111/epi.17200

    work page doi:10.1111/epi.17200 2022

  48. [48]

    Machine learning for detection of interictal epileptiform discharges,

    C. da Silva Lourenço, M. C. Tjepkema -Cloostermans, and M. J. A. M. van Putten, “Machine learning for detection of interictal epileptiform discharges,” Clinical Neurophysiology , vol. 132, no. 7, pp. 1433–1443, Jul. 2021, doi: 10.1016/j.clinph.2021.02.403

    work page doi:10.1016/j.clinph.2021.02.403 2021

  49. [49]

    Moving the field forward: detection of epileptiform abnormalities on scalp electroencephalography using deep learning —clinical application perspectives,

    M. Janmohamed et al. , “Moving the field forward: detection of epileptiform abnormalities on scalp electroencephalography using deep learning —clinical application perspectives,” Brain Commun, vol. 4, no. 5, Sep. 2022, doi: 10.1093/braincomms/fcac218

    work page doi:10.1093/braincomms/fcac218 2022

  50. [50]

    Development of Expert-Level Automated Detection of Epileptiform Discharges During Electroencephalogram Interpretation,

    J. Jing et al., “Development of Expert-Level Automated Detection of Epileptiform Discharges During Electroencephalogram Interpretation,” JAMA Neurol, vol. 77, no. 1, p. 103, Jan. 2020, doi: 10.1001/jamaneurol.2019.3485

    work page doi:10.1001/jamaneurol.2019.3485 2020

  51. [51]

    Spike detection: Inter - reader agreement and a statistical Turing test on a large data set,

    M. L. Scheuer, A. Bagic, and S. B. Wilson, “Spike detection: Inter - reader agreement and a statistical Turing test on a large data set,” Clinical Neurophysiology, vol. 128, no. 1, pp. 243– 250, Jan. 2017, doi: 10.1016/j.clinph.2016.11.005

    work page doi:10.1016/j.clinph.2016.11.005 2017

  52. [52]

    An artificial intelligence- based EEG algorithm for detection of epileptiform EEG discharges: Validation against the diagnostic gold standard,

    F. Fürbass, M. A. Kural, G. Gritsch, M. Hartmann, T. Kluge, and S. Beniczky, “An artificial intelligence- based EEG algorithm for detection of epileptiform EEG discharges: Validation against the diagnostic gold standard,” Clinical Neurophysiology, vol. 131, no. 6, pp. 1174–1179, Jun. 2020, doi: 10.1016/j.clinph.2020.02.032

    work page doi:10.1016/j.clinph.2020.02.032 2020

  53. [53]

    Accurate identification of EEG recordings with interictal epileptiform discharges using a hybrid approach: Artificial intelligence supervised by human experts,

    M. A. Kural et al., “Accurate identification of EEG recordings with interictal epileptiform discharges using a hybrid approach: Artificial intelligence supervised by human experts,” Epilepsia, vol. 63, no. 5, pp. 1064–1073, May 2022, doi: 10.1111/epi.17206

    work page doi:10.1111/epi.17206 2022

  54. [54]

    Automated Interpretation of Clinical Electroencephalograms Using Artificial Intelligence,

    J. Tveit et al. , “Automated Interpretation of Clinical Electroencephalograms Using Artificial Intelligence,” JAMA Neurol, vol. 80, no. 8, p. 805, Aug. 2023, doi: 10.1001/jamaneurol.2023.1645

    work page doi:10.1001/jamaneurol.2023.1645 2023

  55. [55]

    Spreading depression as an innate antiseizure mechanism,

    I. Tamim et al. , “Spreading depression as an innate antiseizure mechanism,” Nat Commun, vol. 12, no. 1, p. 2206, Apr. 2021, doi: 10.1038/s41467-021-22464-x

    work page doi:10.1038/s41467-021-22464-x 2021

  56. [56]

    The Relationship Between Seizures and Spreading Depolarizations in Patients with Severe Traumatic Brain Injury,

    B. Foreman et al. , “The Relationship Between Seizures and Spreading Depolarizations in Patients with Severe Traumatic Brain Injury,” Neurocrit Care, vol. 37, no. S1, pp. 31– 48, Jun. 2022, doi: 10.1007/s12028-022-01441-2

    work page doi:10.1007/s12028-022-01441-2 2022

  57. [57]

    Association of seizures with cortical spreading depression and peri -infarct depolarisations in the acutely injured human brain,

    M. Fabricius et al. , “Association of seizures with cortical spreading depression and peri -infarct depolarisations in the acutely injured human brain,” Clinical Neurophysiology, vol. 119, no. 9, pp. 1973– 1984, Sep. 2008, doi: 10.1016/j.clinph.2008.05.025

    work page doi:10.1016/j.clinph.2008.05.025 1973

  58. [58]

    Dual-path Convolutional Image -Text Embeddings with Instance Loss,

    Z. Zheng, L. Zheng, M. Garrett, Y. Yang, M. Xu, and Y.- D. Shen, “Dual-path Convolutional Image -Text Embeddings with Instance Loss,” ACM Transactions on Multimedia Computing, Communications, and Applications , vol. 16, no. 2, pp. 1– 23, May 2020, doi: 10.1145/3383184

    work page doi:10.1145/3383184 2020

  59. [59]

    Rectifier Nonlinearities Improve Neural Network Acoustic Models,

    A. L. Mass, A. Y. Hannum, and A. Y. Ng, “Rectifier Nonlinearities Improve Neural Network Acoustic Models,” in Proceedings of the 30th International Conference on Machine Learning, Atlanta, Georgia, USA, 2013

    work page 2013

  60. [60]

    Explainable Artificial Intelligence (XAI): What we know and what is left to attain Trustworthy Artificial Intelligence,

    S. Ali et al. , “Explainable Artificial Intelligence (XAI): What we know and what is left to attain Trustworthy Artificial Intelligence,” Information Fusion, vol. 99, p. 101805, Nov. 2023, doi: 10.1016/j.inffus.2023.101805

    work page doi:10.1016/j.inffus.2023.101805 2023

  61. [61]

    Numerical Simulation of Concussive- Generated Cortical Spreading Depolarization to Optimize DC -EEG Electrode Spacing for Noninvasive Visual Detection,

    S. J. Hund et al. , “Numerical Simulation of Concussive- Generated Cortical Spreading Depolarization to Optimize DC -EEG Electrode Spacing for Noninvasive Visual Detection,” Neurocrit Care, vol. 37, no. S1, pp. 67–82, Jun. 2022, doi: 10.1007/s12028-021-01430-x

    work page doi:10.1007/s12028-021-01430-x 2022

  62. [62]

    Adam: A Method for Stochastic Optimization,

    D. P. Kingma and J. Ba, “Adam: A Method for Stochastic Optimization,” Dec. 2014

    work page 2014

  63. [63]

    Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels,

    Z. Zhang and M. R. Sabuncu, “Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels,” May 2018

    work page 2018

  64. [64]

    Clinical correlations of photoparoxysmal responses,

    W. A. Cobb, R. J. Guiloff, and J. Cast, “Breach rhythm: The EEG related to skull defects,” Electroencephalogr Clin Neurophysiol, vol. 47, no. 3, pp. 251– 271, Sep. 1979, doi: 10.1016/0013- 4694(79)90278-5

    work page doi:10.1016/0013- 1979

  65. [65]

    The breach rhythm,

    F. Brigo, R. Cicero, A. Fiaschi, and L. G. Bongiovanni, “The breach rhythm,” Clinical Neurophysiology, vol. 122, no. 11, pp. 2116–2120, Nov. 2011, doi: 10.1016/j.clinph.2011.07.024

    work page doi:10.1016/j.clinph.2011.07.024 2011