Pith. sign in

REVIEW 4 major objections 8 minor 49 references

Reviewed by Pith at T0; open to challenge.

T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →

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

T0 review · glm-5.2

Add the noise back in: brain signal decoders improve when artifacts are used as training data

2026-07-08 01:20 UTC pith:LDBOZFUA

load-bearing objection PNA is a clever inversion of ICA artifact removal into augmentation, with a clean theoretical connection to Jacobian regularization. But the abstract's headline numbers don't match the main results table, and the empirical evidence is thin. the 4 major comments →

arxiv 2607.05165 v1 pith:LDBOZFUA submitted 2026-07-06 cs.LG

Physiological Noise Augmentation Improves Non-Invasive Brain-to-Speech

Benjamin Ballyk , Teyun Kwon , Miran \"Ozdogan , Oiwi Parker Jones This is my paper
classification cs.LG
keywords physiological noise augmentationbrain-to-speech decodingMEGindependent component analysisdata augmentationJacobian regularizationimagined speechbrain-computer interface
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.

Non-invasive brain recordings are dominated by physiological artifacts—eye movements, heartbeat—that standard practice identifies via independent component analysis (ICA) and discards. This paper proposes the opposite: keep those artifact components, scale them to match their empirically observed amplitudes, and remix them back into cleaned brain data as training perturbations. The method, called Physiological Noise Augmentation (PNA), draws a direct analogy to speech recognition systems that add environmental noise to clean audio to build robustness. The central claim is that PNA trains neural decoders to become invariant to artifact directions in input space, and that this invariance can be formalized as anisotropic Jacobian regularization—a penalty on decoder sensitivity specifically along directions where artifacts vary. On MegNIST, a 12,000-trial MEG dataset of imagined digits, PNA combined with 10-trial averaging raised EEGNet decoding accuracy from 73.0% to 76.3% (the abstract reports 77.6%, corresponding to a 4.7-point gain). The paper also shows PNA is complementary to multi-trial averaging: averaging suppresses untracked residual noise while PNA builds invariance to the tracked artifacts, and the two together approximate a structured regularization that targets nuisance directions rather than applying uniform shrinkage.

Core claim

The paper's central mechanism is the inversion of the standard ICA artifact-removal pipeline. Instead of decomposing brain recordings into independent components, identifying artifact-correlated components via reference channels (EOG for ocular, ECG for cardiac), and subtracting them, PNA subtracts them to create clean data but then re-injects scaled copies drawn from the empirical distribution of artifact-to-signal amplitude ratios. This produces biophysically realistic augmented samples that preserve the label while varying the nuisance structure. The theoretical contribution (Proposition 1 and its corollaries) shows that, under multi-trial averaging, this procedure is equivalent in Expect

What carries the argument

PNA (Physiological Noise Augmentation): ICA decomposition → artifact component identification via reference-channel correlation → stochastic rescaling using empirical amplitude ratios → remixing into cleaned data. Theoretical link: PNA ≈ anisotropic Jacobian regularization (Proposition 1), where the regularization matrix is the empirical artifact covariance, penalizing sensitivity along artifact-dominated directions rather than uniformly.

Load-bearing premise

The paper assumes that ICA components correlated with EOG and ECG references are purely task-agnostic—that the brain recording splits cleanly into task signal plus artifact, with no digit-discriminative information leaking into the artifact components. If this separation is imperfect, PNA trains the decoder to become invariant to useful signal, which would work against the intended gains.

What would settle it

Train a simple classifier on the artifact components that PNA identifies and removes. If that classifier achieves above-chance digit classification, the clean-separation assumption is violated and PNA's invariance objective is partly counterproductive.

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

If this is right

  • If PNA generalizes beyond MegNIST to broader vocabularies and multiple subjects, it could reduce the number of trial repetitions a patient must perform at inference time, directly addressing the latency burden that limits clinical viability of non-invasive speech BCIs.
  • The anisotropic regularization interpretation suggests that the choice of which artifacts to track (EOG, ECG, EMG, etc.) determines the regularization geometry—adding more artifact references would shape the decoder's invariance profile more completely.
  • PNA's principle of using domain-specific nuisance structure as augmentation could transfer to other neural recording modalities (fNIRS, intracranial EEG) where reference-tracked noise confounds task signals.
  • The complementarity of PNA and trial averaging implies a trade-off curve: more aggressive artifact augmentation could substitute for some number of trial repetitions, and mapping this curve would quantify the practical repetition savings.

Where Pith is reading between the lines

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

  • The paper validates PNA on a single subject and a 10-class digit task. Whether the 4.7-point gain scales to larger vocabularies (where class boundaries are finer and artifact invariance may be harder to achieve) is untested.
  • If ICA components correlated with EOG/ECG also carry task-relevant neural signal (which the paper does not independently verify), PNA would train the decoder to ignore useful information. A direct test would be to decode digit identity from the discarded artifact components—if above chance, the clean separation assumption is violated.
  • The theoretical result relies on the model being near-convergence (loss gradient ≈ 0) or locally linear in artifact directions. Whether these conditions hold during early training, where most learning happens, is unclear; the regularization interpretation may describe the fine-tuning phase more than the learning phase.

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

4 major / 8 minor

Summary. This paper introduces Physiological Noise Augmentation (PNA), a data augmentation method for non-invasive brain-to-speech decoding. PNA uses ICA to decompose MEG/EEG recordings into task-relevant signals and task-agnostic artifacts (e.g., ocular, cardiac). The artifact components are stochastically scaled and remixed into the cleaned data to generate biophysically realistic, label-preserving training examples. The authors provide a theoretical analysis showing that PNA approximates anisotropic Jacobian regularization under multi-trial averaging, penalizing decoder sensitivity along artifact-dominated directions. Empirically, PNA is evaluated on the MegNIST dataset (single-subject imagined-digit MEG classification) using MLP and EEGNet architectures, comparing against several standard input-level augmentations.

Significance. The paper presents a principled and physiologically grounded approach to data augmentation in neural decoding, which is a meaningful contribution to the non-invasive BCI literature. The theoretical derivation (Proposition 1 and Corollaries 1-2) cleanly connects noise injection to anisotropic Jacobian regularization using standard Taylor expansion techniques, providing a solid theoretical foundation. The method is intuitive and well-motivated. However, the empirical evaluation is limited to a single subject on a constrained 10-class digit classification task, and the headline numerical claims do not match the reported results, which significantly tempers the demonstrated significance.

major comments (4)
  1. Abstract and Section 5 (Discussion): The headline quantitative claim is not supported by the main results table. The abstract states 'PNA with 10-trial averaging improves EEGNet decoding accuracy by 4.7 percentage points (absolute) over training on real data alone' and 'reaching 77.6% decoding accuracy.' However, Table 1 shows the relevant comparison: Raw Only + 10-trial averaging = 73.0% ± 1.4%, and Raw + PNA + 10-trial averaging = 76.3% ± 1.1%. This is a 3.3 pp gap, not 4.7 pp, and 76.3% ≠ 77.6%. The number 77.6% does not appear in Table 1 or Table 2. Figure 5B suggests that higher augmented-to-raw ratios yield higher accuracy, but the main comparison in Table 1 uses a 1:1 ratio (p=0.5). The central quantitative claim must be corrected to match the reported experimental configuration, or the configuration yielding 77.6% must be explicitly identified and made the primary result.
  2. Table 1 and Table 2: The reported improvement for the primary configuration (EEGNet, 10-trial averaging, Raw + PNA vs. Raw Only) is 73.0% ± 1.4% to 76.3% ± 1.1%. With 5 seeds, the combined standard error is approximately 1.78 pp, yielding a z-score of ~1.85. This is marginally non-significant at conventional thresholds (p ≈ 0.06). The paper does not report statistical tests. Given that this 3.3 pp gain is the central empirical finding, the authors should either report significance tests, increase the number of seeds to establish statistical robustness, or explicitly acknowledge the marginal significance of the result.
  3. Table 2: PNA does not consistently outperform simpler baselines. For the MLP with 10-trial averaging, Frequency Shift augmentation achieves 59.6% ± 2.2% compared to PNA's 57.4% ± 2.0%. For EEGNet single-trial, Amplitude Scaling achieves 34.1% ± 0.9% compared to PNA's 33.2% ± 1.1%. The paper's framing in the abstract and discussion emphasizes PNA's superiority, but the evidence in Table 2 shows that PNA is the best method only for the EEGNet + 10-trial averaging configuration. The authors should temper their claims to accurately reflect that PNA's advantage is architecture- and configuration-dependent, or provide a clearer justification for why the EEGNet + averaging setting is the most practically relevant.
  4. Section 3, Eq. (1): The method assumes a clean additive separation X = X_task + X_artifact, where ICA components correlated with EOG/ECG are purely task-agnostic. This is a load-bearing assumption: if ICA artifact components also carry task-relevant neural signal (which is plausible given imperfect separation), then PNA trains the decoder to become invariant to useful information. The paper does not validate that the removed ICA components contain no digit-discriminative information. The authors should discuss this risk and, ideally, verify that the discarded components do not carry task-relevant signal (e.g., by training a classifier on the artifact components alone).
minor comments (8)
  1. Figure 5B: The y-axis label reads 'EEG/glyph1197et' instead of 'EEGNet'. This appears to be a rendering or typographical error.
  2. Figure 9B: Same rendering issue — the y-axis label reads 'EEG/glyph1197et' instead of 'EEGNet'.
  3. Section 3.1 (iii): The text states 'In our experiments, however, we set α_p = 1 for all p to maximize the fidelity of the augmented data.' However, Proposition 1 and its corollaries show that the regularization strength scales with α²/K. If α is always set to 1, the theoretical analysis suggests the regularization effect is entirely determined by K and the empirical artifact covariance. The authors could clarify whether varying α was explored and whether it could serve as a tunable regularization parameter.
  4. Section 4.1: The paper mentions 'We use a 1:1 raw-to-augmented data ratio for simplicity and as supported by Figure 5.' Figure 5B shows EEGNet accuracy continuing to increase up to the 9:1 ratio. The justification for choosing 1:1 is not strongly supported by the figure and could be elaborated upon.
  5. Appendix C, Figure 7 caption: 'intgriguing' should be 'intriguing'.
  6. Section 3.2, Proposition 1: The assumption that E[δ | x̄, y] = 0 is stated, but the paper does not explicitly verify that the sampled artifact perturbations are conditionally mean-zero given the clean input and label. Since artifacts are sampled from donor trials independently of the current trial's label, this seems reasonable, but a brief justification would strengthen the theoretical argument.
  7. Table 1: The 'PNA Only' condition (p=1) shows a substantial performance drop for EEGNet with averaging (62.6% vs. 76.3% for Raw + PNA). This suggests that training exclusively on augmented data is harmful. The authors could briefly discuss why including raw data alongside augmented data is critical.
  8. Section 5 (Discussion): The claim that PNA 'substantially reduces the repetition burden required for high-accuracy decoding' is not directly quantified. The paper shows results at fixed K=10 averaging; it does not demonstrate that PNA allows comparable accuracy at lower K. Figure 4 suggests PNA's benefit is consistent beyond 2 trials, but a direct comparison (e.g., PNA at K=5 vs. Raw at K=10) would substantiate this claim.

Simulated Author's Rebuttal

4 responses · 0 unresolved

We thank the referee for a careful and constructive review. The referee raises four major points: (1) the headline quantitative claim (4.7 pp, 77.6%) does not match Table 1 (3.3 pp, 76.3%); (2) the central empirical result is marginally non-significant without statistical tests; (3) PNA does not consistently outperform simpler baselines across all configurations; and (4) the assumption that ICA artifact components carry no task-relevant signal is unvalidated. We agree with points 1, 2, and 4 and will revise accordingly. On point 3, we partially agree and will temper our claims, while also clarifying why the EEGNet + averaging setting is the practically relevant configuration.

read point-by-point responses

  1. Referee: Abstract and Section 5: The headline quantitative claim (4.7 pp, 77.6%) does not match Table 1 (3.3 pp, 76.3%). The number 77.6% does not appear in any table. Figure 5B suggests higher augmented-to-raw ratios yield higher accuracy, but the main comparison uses 1:1 (p=0.5).

    Authors: The referee is correct. The abstract and discussion cite a 4.7 pp improvement to 77.6%, but Table 1 reports 73.0% → 76.3% (a 3.3 pp gap) at the 1:1 ratio (p=0.5). The 77.6% figure and the 4.7 pp gap correspond to a higher augmented-to-raw ratio shown in Figure 5B, which is not the primary configuration reported in Table 1. This is an inconsistency between our headline claims and our main results table, and it arose from inadvertently reporting the best Figure 5B data point in the abstract rather than the Table 1 configuration. We will correct the abstract and discussion to state the Table 1 result: a 3.3 pp improvement from 73.0% to 76.3% with EEGNet under 10-trial averaging at p=0.5. We will also explicitly identify the configuration yielding 77.6% (higher augmented-to-raw ratio from Figure 5B) and clarify its relationship to the primary 1:1 comparison. revision: yes

  2. Referee: Table 1 and Table 2: The 3.3 pp improvement (73.0% ± 1.4% to 76.3% ± 1.1%) with 5 seeds yields a z-score of ~1.85, marginally non-significant (p ≈ 0.06). No statistical tests are reported.

    Authors: The referee's calculation is correct, and we agree that statistical testing is necessary for the central empirical claim. With 5 seeds and the reported standard errors, the improvement is marginally non-significant at conventional thresholds. We will address this in two ways: (1) we will increase the number of seeds to at least 10 to provide a more robust estimate of the effect and its significance, and (2) we will report paired statistical tests (e.g., paired t-test or Wilcoxon signed-rank test across seeds) for all primary comparisons. If the result remains marginally significant after increasing seeds, we will explicitly acknowledge this in the revised manuscript rather than overstating the finding. revision: yes

  3. Referee: Table 2: PNA does not consistently outperform simpler baselines. Frequency Shift beats PNA for MLP + 10-trial averaging (59.6% vs 57.4%), and Amplitude Scaling beats PNA for EEGNet single-trial (34.1% vs 33.2%). PNA is best only for EEGNet + 10-trial averaging.

    Authors: We partially agree. The referee correctly identifies that PNA does not dominate across all architecture and configuration combinations. We will temper the framing in the abstract and discussion to accurately reflect that PNA's advantage is concentrated in the EEGNet + 10-trial averaging setting, rather than claiming blanket superiority. However, we also wish to justify why this setting is the most practically relevant: multi-trial averaging is standard practice in non-invasive BCI for SNR enhancement, and EEGNet is a widely used architecture for M/EEG decoding. The single-trial regime is fundamentally SNR-limited (as our own results show—no method meaningfully exceeds chance-level performance for MLP), so it is not the deployment scenario where augmentation strategies are expected to provide meaningful gains. We will add this justification to the discussion and present the cross-configuration results transparently, acknowledging where simpler augmentations are competitive or superior. revision: partial

  4. Referee: Section 3, Eq. (1): The clean additive separation X = X_task + X_artifact assumes ICA artifact components are purely task-agnostic. If artifact components carry task-relevant neural signal, PNA trains the decoder to become invariant to useful information. This is not validated.

    Authors: The referee raises a valid and important concern. The assumption that ICA components correlated with EOG/ECG references are purely task-agnostic is load-bearing for PNA's correctness, and imperfect ICA separation could indeed mean that some task-relevant neural signal is captured in artifact components. We currently do not validate that discarded components carry no digit-discriminative information. We will address this in the revision by training a classifier (EEGNet) on the artifact components alone to test whether they contain above-chance digit-discriminative signal. If the artifact components do carry some task-relevant information, we will discuss this as a limitation and potential explanation for why PNA's gains are modest and configuration-dependent. We will also add a discussion of this risk to Section 3 alongside Equation (1), noting that PNA's effectiveness depends on the quality of artifact-relevant separation, and that future work could use more sophisticated component classification to mitigate this issue. revision: yes

Circularity Check

0 steps flagged

No circularity found; derivation is self-contained with standard mathematical machinery

full rationale

The paper's theoretical result (Proposition 1 and its corollaries) is a standard multivariate Taylor expansion of the loss under Gaussian perturbation, following the classical Bishop (1995) framework. The proof in Appendix B proceeds by expanding logits to second order, substituting into the loss, taking expectation, and applying trace identities — all standard calculus with no self-citation chain. The key identity E[δ^T A δ] = Tr(A Σ_δ) is a basic statistical fact, not a result from the authors' prior work. The empirical augmentation pipeline (sampling artifact-to-clean ratios from training data) is legitimate augmentation practice: it matches test-time artifact statistics, which is the standard goal of domain-matched augmentation. The α=1 setting is justified empirically via Figure 2, not by definition. The ICA decomposition uses FastICA (Hyvärinen 1999) and MNE-Python — external tools. No step in the derivation chain reduces to its own inputs by construction. The numerical discrepancy between the abstract (77.6%, 4.7 pp) and Table 1 (76.3%, 3.3 pp) is a correctness/reporting concern, not a circularity issue — it does not involve any step being equivalent to its inputs by definition.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 0 invented entities

No new physical entities, particles, forces, or dimensions are introduced.

free parameters (4)
  • α_p (per-artifact scaling) = 1.0 for all p
    Set to 1 to match empirical artifact distribution; could be tuned but is fixed in experiments.
  • τ_p (correlation thresholds for artifact component selection) = Not explicitly stated
    Thresholds for |ρ(ŝ_i, ϕ_p)| used to identify artifact components; values not reported in main text.
  • ε (numerical stability constant) = Small positive, not specified
    Added to denominators for numerical stability; exact value not given.
  • p (augmentation probability) = 0.5
    Probability of applying PNA during training; chosen by the authors, supported by Figure 5 sweep.
axioms (4)
  • domain assumption X = X_task + X_artifact with clean additive separation
    Section 3, Eq. 1. Assumes neural signal and artifacts are additively separable, which is standard but imperfect.
  • standard math ICA decomposes recordings into statistically independent sources
    Section 3.1(ii), Appendix A.1. Standard ICA assumption (Bell & Sejnowski 1995, Comon 1994).
  • domain assumption Artifact components identified by EOG/ECG correlation are task-agnostic
    Section 3.1(ii). Assumes components correlated with reference channels contain no task-relevant signal. Not independently validated.
  • ad hoc to paper Decoder is locally linear in artifact directions OR loss gradient ≈ 0
    Proposition 1, conditions (a)/(b). Required to simplify the regularization result to a pure Jacobian penalty. Condition (b) is argued to hold late in training but not verified.

pith-pipeline@v1.1.0-glm · 18483 in / 3586 out tokens · 194368 ms · 2026-07-08T01:20:12.546458+00:00 · methodology

0 comments
read the original abstract

Non-invasive brain-to-speech decoding aims to restore communication to patients suffering from neurodegenerative disease, without the risks of neurosurgery. Existing MEG- and EEG-based methods, while scalable, continue to suffer from high word error rates driven by relatively low signal-to-noise ratios compared to invasive recordings. We propose physiological noise augmentation (PNA), a data augmentation method that explicitly trains decoders to become invariant to task-agnostic artifacts (e.g. ocular and cardiac activity). PNA draws inspiration from automatic speech recognition systems, where environmental noise (e.g. dogs barking, city traffic) is added to clean speech to improve robustness. Analogously, we decompose brain recordings into clean data and noise artifacts using independent component analysis (ICA), before scaling and remixing to generate biophysically realistic, label-preserving training examples. We show that PNA approximates anisotropic regularization, penalizing decoder sensitivity along artifact-dominated directions. On MegNIST, a 12k-trial imagined-digit MEG dataset, PNA with 10-trial averaging improves EEGNet decoding accuracy by 4.7 percentage points (absolute) over training on real data alone. Our results suggest that artifact-aware augmentation and trial averaging are complementary tools for improving robustness in non-invasive speech BCIs.

Figures

Figures reproduced from arXiv: 2607.05165 by Benjamin Ballyk, Miran \"Ozdogan, Oiwi Parker Jones, Teyun Kwon.

Figure 1
Figure 1. Figure 1: Overview of physiological noise augmentation. (A) Raw brain recordings (Xraw) and reference channels (e.g. ocular ϕeog and cardiac ϕecg) are collected. (B) Independent component analysis (ICA) decomposes Xraw into independent components, which are correlated with references to identify artifact components. (C) Artifact components are projected back to sensor space, stochas￾tically scaled using their empiri… view at source ↗
Figure 2
Figure 2. Figure 2: Stochastic amplitude sampling calibrates PNA-augmented MEG embeddings to the dis￾tribution of raw data. (Top) Empirical artifact-to-clean amplitude ratio distributions (zeros removed). (Bottom) PCA fit to 15-trial-averaged raw (red) and clean (black) embeddings of imagined digits; augmented samples (blue, gold centroid) are projected at varying scalings α. Augmented embeddings interpolate between clusters … view at source ↗
Figure 3
Figure 3. Figure 3: t-SNE embeddings of single-trial (left) and 15-trial averaged (right) intra-patient MEG recordings of imagined digits (0–9). Averaging is performed with resampling to preserve dataset cardinality; t-SNE uses perplexity = 30. imagined digit classes. In contrast, clear class-specific clusters emerge after 15-trial averaging. We repeat this analysis in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: MLP decoding accuracy uplift from PNA augmentation and associated t-SNE em￾beddings at various levels of averaging. The top plot uses 10k training trials for raw-only runs (red) and an additional 10k augmented trials for raw + augmented runs (blue), both over 5 seeds. The bottom plots show t-SNE embeddings of raw (red triangles) and augmented (blue dots) data at each level of averaging [PITH_FULL_IMAGE:fi… view at source ↗
Figure 5
Figure 5. Figure 5: Decoding accuracy scales with PNA samples. Models are trained on a fixed budget of 10k raw trials; the x-axis represents the ratio of augmented-to-raw data, ranging from 0 (baseline, vertical dotted line) to 90k (9 : 1 ratio). Shaded regions denote 95% confidence intervals (n = 5 independent seeds); all runs use 10-trial averaging [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: ICA artifact component selection for EOG and ECG signals. The topographies captured [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: UMAP embeddings of intra-patient MEG recordings for imagined digits (0–9), shown for [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: PCA and t-SNE embeddings of raw, augmented (ocular and cardiac), and clean data at [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Decoding accuracy scales with PNA samples (averaging without resampling). Models are trained on a fixed budget of 10k raw trials (1k post-averaging); the x-axis represents the ratio of augmented-to-raw data, ranging from 0 (baseline, vertical dotted line) to 90k (9 : 1 ratio; 9k post-averaging). Shaded regions denote 95% confidence intervals (n = 5 independent seeds); all runs 10-trial averaging. 0 1:4 1:2… view at source ↗
Figure 10
Figure 10. Figure 10: Single-Trial Decoding accuracy under various PNA data infusions. Models are trained on a fixed budget of 10k raw trials; the x-axis represents the ratio of augmented-to-raw data, ranging from 0 (baseline, vertical dotted line) to 90k (9 : 1 ratio). Shaded regions denote 95% confidence intervals (n = 5 independent seeds). For “Raw + Aug.”, we concatenate varying amounts of augmented samples to 10,000 raw s… 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

49 extracted references · 49 canonical work pages · 6 internal anchors

  1. [1]

    Speech synthesis from neural decoding of spoken sentences.Nature, 568(7753):493–498, 2019

    Gopala K Anumanchipalli, Josh Chartier, and Edward F Chang. Speech synthesis from neural decoding of spoken sentences.Nature, 568(7753):493–498, 2019

    work page 2019

  2. [2]

    Bell and Terrence J

    Anthony J. Bell and Terrence J. Sejnowski. An information-maximization approach to blind separation and blind deconvolution.Neural Computation, 7(6):1129–1159, 1995. doi: 10.1162/ neco.1995.7.6.1129

    work page 1995

  3. [3]

    Birbaumer, N

    N. Birbaumer, N. Ghanayim, T. Hinterberger, I. Iversen, B. Kotchoubey, A. Kübler, J. Perel- mouter, E. Taub, and H. Flor. A spelling device for the paralysed.Nature, 398(6725):297–298, Mar 1999. ISSN 1476-4687. doi: 10.1038/18581. URL https://doi.org/10.1038/18581

    work page doi:10.1038/18581 1999

  4. [4]

    Training with noise is equivalent to tikhonov regularization.Neural computa- tion, 7(1):108–116, 1995

    Chris M Bishop. Training with noise is equivalent to tikhonov regularization.Neural computa- tion, 7(1):108–116, 1995

    work page 1995

  5. [5]

    Card, Maitreyee Wairagkar, Carrina Iacobacci, Xianda Hou, Tyler Singer-Clark, Francis R

    Nicholas S. Card, Maitreyee Wairagkar, Carrina Iacobacci, Xianda Hou, Tyler Singer-Clark, Francis R. Willett, Erin M. Kunz, Chaofei Fan, Maryam Vahdati Nia, Darrel R. Deo, Aparna Srinivasan, Eun Young Choi, Matthew F. Glasser, Leigh R. Hochberg, Jaimie M. Henderson, Kiarash Shahlaie, Sergey D. Stavisky, and David M. Brandman. An accurate and rapidly calib...

    work page doi:10.1056/nejmoa2314132 2024

  6. [6]

    Independent component analysis, a new concept?Signal processing, 36(3): 287–314, 1994

    Pierre Comon. Independent component analysis, a new concept?Signal processing, 36(3): 287–314, 1994

    work page 1994

  7. [7]

    Towards decoding individual words from non-invasive brain recordings.Nature Communications, 16(1):10521, Nov 2025

    Stéphane d’Ascoli, Corentin Bel, Jérémy Rapin, Hubert Banville, Yohann Benchetrit, Christophe Pallier, and Jean-Rémi King. Towards decoding individual words from non-invasive brain recordings.Nature Communications, 16(1):10521, Nov 2025. ISSN 2041-1723. doi: 10.1038/ s41467-025-65499-0

    work page 2025

  8. [8]

    A summation technique for the detection of small evoked potentials

    George D Dawson. A summation technique for the detection of small evoked potentials. Electroencephalography & clinical neurophysiology, 1954

    work page 1954

  9. [9]

    Megconformer: Conformer-based meg decoder for robust speech and phoneme classification.arXiv e-prints, page arXiv:2512.01443, December 2025

    Xabier de Zuazo, Ibon Saratxaga, and Eva Navas. Megconformer: Conformer-based meg decoder for robust speech and phoneme classification.arXiv e-prints, page arXiv:2512.01443, December 2025. doi: 10.48550/arXiv.2512.01443

    work page doi:10.48550/arxiv.2512.01443 2025

  10. [10]

    Decoding speech perception from non-invasive brain recordings.Nature Machine Intelligence, 5(10):1097–1107, Oct 2023

    Alexandre Défossez, Charlotte Caucheteux, Jérémy Rapin, Ori Kabeli, and Jean-Rémi King. Decoding speech perception from non-invasive brain recordings.Nature Machine Intelligence, 5(10):1097–1107, Oct 2023. ISSN 2522-5839. doi: 10.1038/s42256-023-00714-5

    work page doi:10.1038/s42256-023-00714-5 2023

  11. [11]

    Talking off the top of your head: toward a mental prosthesis utilizing event-related brain potentials.Electroencephalography and clinical Neurophysiology, 70(6):510–523, 1988

    Lawrence Ashley Farwell and Emanuel Donchin. Talking off the top of your head: toward a mental prosthesis utilizing event-related brain potentials.Electroencephalography and clinical Neurophysiology, 70(6):510–523, 1988

    work page 1988

  12. [12]

    Valery L Feigin, Benjamin A Stark, Catherine Owens Johnson, Gregory A Roth, Catherine Bisignano, Gdiom Gebreheat Abady, Mitra Abbasifard, Mohsen Abbasi-Kangevari, Foad Abd- Allah, Vida Abedi, et al. Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the global burden of disease study 2019.The Lancet ...

    work page 1990

  13. [13]

    GBDTB. Global, regional, and national burden of traumatic brain injury and spinal cord injury, 1990-2016: a systematic analysis for the global burden of disease study 2016.Lancet Neurology, 18(1):56–87, 2019

    work page 1990

  14. [14]

    Meg and eeg data analysis with mne-python.Frontiers in Neuroinformatics, 7:267, 2013

    Alexandre Gramfort, Martin Luessi, Eric Larson, Denis A Engemann, Daniel Strohmeier, Christian Brodbeck, Roman Goj, Mainak Jas, Teon Brooks, Lauri Parkkonen, et al. Meg and eeg data analysis with mne-python.Frontiers in Neuroinformatics, 7:267, 2013

    work page 2013

  15. [15]

    Data augmentation for deep neural networks model in eeg classification task: a review.Frontiers in Human Neuroscience, 15:765525, 2021

    Chao He, Jialu Liu, Yuesheng Zhu, and Wencai Du. Data augmentation for deep neural networks model in eeg classification task: a review.Frontiers in Human Neuroscience, 15:765525, 2021. 10

    work page 2021

  16. [16]

    Fast and robust fixed-point algorithms for independent component analysis

    Aapo Hyvärinen. Fast and robust fixed-point algorithms for independent component analysis. IEEE Transactions on Neural Networks, 10(3):626–634, 1999. doi: 10.1109/72.761722

    work page doi:10.1109/72.761722 1999

  17. [17]

    Independent component analysis: algorithms and applications

    Aapo Hyvärinen and Erkki Oja. Independent component analysis: algorithms and applications. Neural Networks, 13(4-5):411–430, 2000. doi: 10.1016/S0893-6080(00)00026-5

    work page doi:10.1016/s0893-6080(00)00026-5 2000

  18. [18]

    Unlocking Non-Invasive Brain-to-Text

    Dulhan Jayalath, Gilad Landau, and Oiwi Parker Jones. Unlocking non-invasive brain-to-text. arXiv e-prints, page arXiv:2505.13446, May 2025. doi: 10.48550/arXiv.2505.13446

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2505.13446 2025

  19. [19]

    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 internal anchor Pith review Pith/arXiv arXiv 2024

  20. [20]

    Removing electroencephalographic artifacts by blind source separation.Psychophysiology, 37(2):163–178, 2000

    Tzyy-Ping Jung, Scott Makeig, Colin Humphries, Te-Won Lee, Martin J Mckeown, Vicente Iragui, and Terrence J Sejnowski. Removing electroencephalographic artifacts by blind source separation.Psychophysiology, 37(2):163–178, 2000

    work page 2000

  21. [21]

    A study on data augmentation of reverberant speech for robust speech recognition

    Tom Ko, Vijayaditya Peddinti, Daniel Povey, Michael L Seltzer, and Sanjeev Khudanpur. A study on data augmentation of reverberant speech for robust speech recognition. In2017 IEEE international conference on acoustics, speech and signal processing (ICASSP), pages 5220–5224. IEEE, 2017

    work page 2017

  22. [22]

    EEGNet: A compact convolutional neural network for EEG -based brain–computer interfaces,

    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, July 2018. doi: 10.1088/1741-2552/aace8c

    work page doi:10.1088/1741-2552/aace8c 2018

  23. [23]

    Leuthardt, Gerwin Schalk, Jarod Roland, Andrew Rouse, and Daniel Moran

    Eric C. Leuthardt, Gerwin Schalk, Jarod Roland, Andrew Rouse, and Daniel Moran. Evolution of brain–computer interfaces: Going beyond classic motor physiology.Frontiers in Neuroscience, 15:599549, 2021. doi: 10.3389/fnins.2021.599549

    work page doi:10.3389/fnins.2021.599549 2021

  24. [24]

    Decoupled Weight Decay Regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. InInternational Conference on Learning Representations (ICLR), 2019. URL https://openreview.net/ forum?id=Bkg6RiCqY7. arXiv:1711.05101

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  25. [25]

    Channel drop out: A simple way to prevent cnn from overfitting in motor imagery based bci

    Jing Luo, Yaojie Wang, Rong Xu, Guangming Liu, Xiaofan Wang, and Yijing Gong. Channel drop out: A simple way to prevent cnn from overfitting in motor imagery based bci. In International Conference of Pioneering Computer Scientists, Engineers and Educators, pages 443–452. Springer, 2021

    work page 2021

  26. [26]

    Crone, Jochem Rieger, Gerwin Schalk, Robert T

    Stéphanie Martin, Peter Brunner, Chris Holdgraf, Hans-Jochen Heinze, Nathan E. Crone, Jochem Rieger, Gerwin Schalk, Robert T. Knight, and Brian N. Pasley. Decoding spectrotempo- ral features of overt and covert speech from the human cortex.Frontiers in Neuroengineering, 7:14, 2014. doi: 10.3389/fneng.2014.00014

    work page doi:10.3389/fneng.2014.00014 2014

  27. [27]

    Millán, Gerwin Schalk, Robert T

    Stephanie Martin, Peter Brunner, Iñaki Iturrate, José del R. Millán, Gerwin Schalk, Robert T. Knight, and Brian N. Pasley. Word pair classification during imagined speech using direct brain recordings.Scientific Reports, 6:25803, 2016. doi: 10.1038/srep25803

    work page doi:10.1038/srep25803 2016

  28. [28]

    Millán, Robert T

    Stephanie Martin, Iñaki Iturrate, José del R. Millán, Robert T. Knight, and Brian N. Pasley. Decoding inner speech using electrocorticography: Progress and challenges toward a speech prosthesis.Frontiers in Neuroscience, V olume 12 - 2018, 2018. ISSN 1662- 453X. doi: 10.3389/fnins.2018.00422. URL https://www.frontiersin.org/journals/ neuroscience/articles...

    work page doi:10.3389/fnins.2018.00422 2018

  29. [29]

    UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

    Leland McInnes, John Healy, and James Melville. Umap: Uniform manifold approximation and projection for dimension reduction.arXiv preprint arXiv:1802.03426, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  30. [30]

    Moses, Sean L

    David A. Moses, Sean L. Metzger, Jessie R. Liu, Gopala K. Anumanchipalli, Joseph G. Makin, Pengfei F. Sun, Josh Chartier, Maximilian E. Dougherty, Patricia M. Liu, Gary M. Abrams, Adelyn Tu-Chan, Karunesh Ganguly, and Edward F. Chang. Neuroprosthesis for decoding speech in a paralyzed person with anarthria.New England Journal of Medicine, 385(3):217–227,

    work page

  31. [31]

    doi: 10.1056/NEJMoa2027540. 11

    work page doi:10.1056/nejmoa2027540

  32. [32]

    Decoding covert speech from eeg-a comprehensive review.Frontiers in Neuroscience, V olume 15 - 2021, 2021

    Jerrin Thomas Panachakel and Angarai Ganesan Ramakrishnan. Decoding covert speech from eeg-a comprehensive review.Frontiers in Neuroscience, V olume 15 - 2021, 2021. ISSN 1662- 453X. doi: 10.3389/fnins.2021.642251. URL https://www.frontiersin.org/journals/ neuroscience/articles/10.3389/fnins.2021.642251

    work page doi:10.3389/fnins.2021.642251 2021

  33. [33]

    SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

    Daniel S. Park, William Chan, Yu Zhang, Chung-Cheng Chiu, Barret Zoph, Ekin D. Cubuk, and Quoc V . Le. Specaugment: A simple data augmentation method for automatic speech recognition.arXiv e-prints, page arXiv:1904.08779, April 2019. doi: 10.48550/arXiv.1904. 08779

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1904 1904

  34. [34]

    The global burden of motor neuron disease: an analysis of the 2019 global burden of disease study.Frontiers in Neurology, 13:864339, 2022

    Jin Park, Jee-Eun Kim, and Tae-Jin Song. The global burden of motor neuron disease: an analysis of the 2019 global burden of disease study.Frontiers in Neurology, 13:864339, 2022

    work page 2019

  35. [35]

    Barbour, Eric C

    Xiaomei Pei, Dennis L. Barbour, Eric C. Leuthardt, and Gerwin Schalk. Decoding vowels and consonants in spoken and imagined words using electrocorticographic signals in humans. Journal of Neural Engineering, 8(4):046028, 2011. doi: 10.1088/1741-2560/8/4/046028

    work page doi:10.1088/1741-2560/8/4/046028 2011

  36. [36]

    Iclabel: An automated elec- troencephalographic independent component classifier, dataset, and website.NeuroImage, 198: 181–197, 2019

    Luca Pion-Tonachini, Ken Kreutz-Delgado, and Scott Makeig. Iclabel: An automated elec- troencephalographic independent component classifier, dataset, and website.NeuroImage, 198: 181–197, 2019

    work page 2019

  37. [37]

    Imagined speech can be decoded from low-and cross-frequency intracranial eeg features.Nature communications, 13 (1):48, 2022

    Timothée Proix, Jaime Delgado Saa, Andy Christen, Stephanie Martin, Brian N Pasley, Robert T Knight, Xing Tian, David Poeppel, Werner K Doyle, Orrin Devinsky, et al. Imagined speech can be decoded from low-and cross-frequency intracranial eeg features.Nature communications, 13 (1):48, 2022

    work page 2022

  38. [38]

    MUSAN: A Music, Speech, and Noise Corpus

    David Snyder, Guoguo Chen, and Daniel Povey. Musan: A music, speech, and noise corpus. arXiv preprint arXiv:1510.08484, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  39. [39]

    Samuel Sutton, Margery Braren, Joseph Zubin, and E. R. John. Evoked-potential correlates of stimulus uncertainty.Science, 150(3700):1187–1188, 1965. doi: 10.1126/science.150.3700

    work page doi:10.1126/science.150.3700 1965

  40. [40]

    URLhttps://www.science.org/doi/abs/10.1126/science.150.3700.1187

    work page doi:10.1126/science.150.3700.1187

  41. [41]

    Visualizing data using t-sne.Journal of machine learning research, 9(11), 2008

    Laurens Van der Maaten and Geoffrey Hinton. Visualizing data using t-sne.Journal of machine learning research, 9(11), 2008

    work page 2008

  42. [42]

    Indepen- dent component approach to the analysis of eeg and meg recordings.IEEE transactions on biomedical engineering, 47(5):589–593, 2002

    Ricardo Vigário, Jaakko Sarela, Veikko Jousmiki, Matti Hamalainen, and Erkki Oja. Indepen- dent component approach to the analysis of eeg and meg recordings.IEEE transactions on biomedical engineering, 47(5):589–593, 2002

    work page 2002

  43. [43]

    Time–frequency transform based eeg data augmentation for brain–computer interfaces.Knowledge-Based Systems, 311:113074,

    Ziwei Wang, Siyang Li, Xiaoqing Chen, and Dongrui Wu. Time–frequency transform based eeg data augmentation for brain–computer interfaces.Knowledge-Based Systems, 311:113074,

    work page

  44. [44]

    doi: https://doi.org/10.1016/j.knosys.2025.113074

    ISSN 0950-7051. doi: https://doi.org/10.1016/j.knosys.2025.113074

    work page doi:10.1016/j.knosys.2025.113074 2025

  45. [45]

    Willett, Erin M

    Francis R. Willett, Erin M. Kunz, Chaofei Fan, Donald T. Avansino, Guy H. Wilson, Eun Young Choi, Foram Kamdar, Matthew F. Glasser, Leigh R. Hochberg, Shaul Druckmann, Krishna V . Shenoy, and Jaimie M. Henderson. A high-performance speech neuroprosthesis.Nature, 620 (7976):1031–1036, 2023. doi: 10.1038/s41586-023-06377-x

    work page doi:10.1038/s41586-023-06377-x 2023

  46. [46]

    Automatic classification of artifactual ica-components for artifact removal in eeg signals.Behavioral and brain functions, 7(1):30, 2011

    Irene Winkler, Stefan Haufe, and Michael Tangermann. Automatic classification of artifactual ica-components for artifact removal in eeg signals.Behavioral and brain functions, 7(1):30, 2011

    work page 2011

  47. [47]

    Wolpaw, Niels Birbaumer, Dennis J

    Jonathan R. Wolpaw, Niels Birbaumer, Dennis J. McFarland, Gert Pfurtscheller, and Theresa M. Vaughan. Brain-computer interfaces for communication and control.Clinical Neurophysiology, 113(6):767–791, June 2002. doi: 10.1016/S1388-2457(02)00057-3. 12 A Additional Background A.1 Independent Component Analysis Independent component analysis(ICA) is modeled o...

    work page doi:10.1016/s1388-2457(02)00057-3 2002

  48. [48]

    While simple, this reduces the effective training set size by a factor of K, limiting input variability

    Averaging without resampling:Trials are partitioned into disjoint groups of size K and averaged. While simple, this reduces the effective training set size by a factor of K, limiting input variability

    work page

  49. [49]

    Raw + Aug

    Averaging with resampling:Subsets of size K are randomly drawn (without replacement within each subset) from the full training set. For a dataset of size N, there are N K possible subsets, a combinatorial explosion (when min(K, N−K)>>0 ) that allows for massive data augmentation while preserving the original dataset’s coverage. In this work, we mainly emp...

    work page