Pith. sign in

REVIEW 3 major objections 5 minor 35 references

Decoding Speaker-Normalized Pitch from EEG for Mandarin Perception

T0 review · 3 major / 5 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 Mandarin speech perception encodes pitch relative to the speaker, and the normalization happens within an isolated syllable.

desk verdict The neural claim is plausible but the decoding comparison is not yet convincing: the normalized condition may be handed speaker means for free at evaluation time, so the 32.7% advantage is not established. read the letter →

arxiv 2505.19626 v1 pith:TJBMIY6Y submitted 2025-05-26 cs.SD eess.AS

Jiaxin Chen , Yiming Wang , Ziyu Zhang , Jiayang Han , Yin-Long Liu , Rui Feng , Xiuyuan Liang , Zhen-Hua Ling
show 1 more author
Jiahong Yuan
This is my paper · ORCID
classification cs.SDeess.AS
keywords EEGdecodingMandarintonesspeakernormalizationrelativepitchcontourCE-ViViTspeechperceptionfundamentalfrequency
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

Using a new EEG dataset in which listeners heard Mandarin monosyllables from one speaker and from six speakers, this paper asks whether the brain tracks absolute pitch contours or pitch contours normalized to each speaker's range. A regression model called CE-ViViT reconstructed pitch contours directly from the EEG, and across all tested systems the speaker-normalized contours decoded more accurately whenever multiple speakers were present, with an average MAE improvement of 32.7%. On the single-speaker session normalized and raw pitch decoded equally well, which the authors use to rule out the possibility that the gain comes only from a smaller dynamic range. The paper concludes that Mandarin perception relies on speaker-normalized relative pitch rather than absolute pitch, and that the normalization happens at the phoneme level during tonal perception.

What carries the argument

The object that carries the argument is the speaker-normalized pitch contour, defined here as the fundamental frequency of each syllable converted to a z-score with that speaker's mean and standard deviation. The tool built to read this contour out of brain activity is CE-ViViT, a regression model that embeds raw EEG via parallel temporal, dilation, and channel convolutional blocks refined by ScConv, fuses the branches with DANE, encodes global relations with a two-stage Video Vision Transformer, and maps the representation to the F0 curve through a pitch head trained with mean squared error. Comparing how accurately normalized versus raw contours are decoded in a one-speaker and a six-speaker design is the experiment that does the work.

What would settle it

A concrete result that would settle the claim would be a multi-speaker decoding study in which raw-pitch targets generalize to held-out speakers as well as normalized-pitch targets do, or in which the normalized-pitch advantage disappears when the z-score is replaced by a different per-speaker transform such as log-frequency centering. Either outcome would undermine the conclusion that the brain encodes speaker-normalized relative pitch at the phoneme level.

Watch

Extended reading notes

Core claim

The paper's central claim is that when a listener hears a Mandarin syllable, the auditory system encodes the pitch contour relative to the speaker's own pitch range, and this normalization is complete at the phoneme level, before any sentence context is available. The evidence is a decoding contrast: on the six-speaker session, the proposed CE-ViViT and the three comparison systems all reconstructed per-speaker z-scored pitch contours with lower error than raw F0 contours, while on the single-speaker session the two targets were decoded equally well. Because the single-speaker result shows that normalization of the target alone does not help, the multi-speaker advantage is interpreted as neural encoding of relative pitch. The authors state the conclusion directly: Mandarin perception relies on speaker-normalized relative pitch rather than absolute pitch, and the normalization process occurs at the phoneme level during tonal perception.

Load-bearing premise

The interpretation depends on the assumption that the brain's speaker normalization is exactly the per-speaker z-score of fundamental frequency; if the brain normalizes pitch by some other rule, the improved decoding of z-scored contours would not demonstrate relative-pitch encoding.

Editorial extensions

If this is right

  • Decoders for tonal speech brain-computer interfaces should target speaker-relative pitch contours, because that is the representation the EEG signal appears to carry.
  • Accounts of tone perception that treat absolute F0 as the perceptual feature omit the normalization step that the decoding results place inside the phoneme.
  • Single-speaker EEG studies cannot separate absolute from relative pitch encoding, so multi-speaker designs are necessary for this question.
  • The phoneme-level finding means speaker normalization does not need sentence context, which constrains where in the auditory pathway it could occur.
  • The average 32.7% MAE improvement on the six-speaker session gives later studies a concrete benchmark for detecting speaker-normalized pitch in non-invasive recordings.

Reading between the lines

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

  • The authors do not test whether the brain's transform is exactly a z-score; identical decoding comparisons using other per-speaker transformations, such as rank normalization or log-frequency centering, would show which code the neural response actually matches.
  • If the relative-pitch claim is right, a decoder trained on several speakers should transfer normalized-pitch decoding to a new speaker better than it transfers raw-pitch decoding, which is a testable extension of the reported result.
  • The monosyllabic design leaves open whether sentence context adjusts the normalization; the authors flag sentence-level context as future work, and one would predict that context either sharpens or overrides the phoneme-level normalization.
  • The result suggests hearing devices for tonal languages could preserve speaker-relative contour shape rather than exact frequencies, since the brain appears to read pitch that way.
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 / 5 minor

Summary. The paper presents a new EEG dataset of Mandarin monosyllable perception (29 listeners, two stimulus sessions: one speaker and six speakers) and proposes a convolutional embedding plus Video Vision Transformer model (CE-ViViT) to regress fundamental-frequency (F0) contours from EEG. The authors compare decoding of raw F0 with decoding of per-speaker z-score normalized F0, reporting that normalized pitch is decoded more accurately than raw pitch in the six-speaker session (32.7% average MAE improvement) while no clear advantage appears in the single-speaker session. From this they conclude that Mandarin tonal perception relies on speaker-normalized relative pitch and that normalization occurs at the phoneme level. The model is also ablated to justify its components.

Significance. If the central claim were established, the paper would extend invasive ECoG evidence of speaker-normalized relative pitch encoding (Tang et al., 2017) to non-invasive EEG and to isolated monosyllables, which is a meaningful contribution for both neuroscience and auditory EEG decoding. The paper's strengths are the collection of a new auditory EEG dataset with controlled speaker/tone/phoneme variation, the use of multiple SOTA baseline models, and a clearly described architecture with an ablation study. However, the main conclusion currently rests on a metric-comparison asymmetry and on aggregate numbers without variance or significance testing; the significance of the paper is therefore contingent on a careful re-analysis.

major comments (3)
  1. [Section 4.1 and Section 3.3] The central comparison is not fair as implemented. The text states that speaker-normalized pitch contours are 'rescaled to the original raw pitch contours’ scale using the speaker’s mean and variance during metric calculation', while raw predictions are evaluated directly. This injects oracle per-speaker statistics into the normalized condition. A trivial normalized decoder that always outputs the z-score mean (0) becomes, after rescaling, a per-trial prediction of that speaker's mean F0, with MAE = E|F0 − μ_s|; the corresponding raw trivial baseline (global mean) has MAE = E|F0 − μ_g|. The gap between these baselines is largely between-speaker F0 variability and exists even if the EEG contains no pitch information. This explains the observed pattern: no advantage in session 1 (one speaker) and a large advantage in session 2 (six speakers). The optional identity vector does not remove this asymmetry, because the raw condition must learn speaker means from data while the normalized condition is given them for free during metric computation. The authors should report trivial speaker-mean and global-mean baselines, and/or rescale both conditions using training-set speaker statistics, and/or evaluate all predictions on a common normalized scale.
  2. [Tables 1 and 2, Section 3.2] No variance, confidence interval, or significance test is reported for any table, yet the conclusion depends on 'no significant difference' in session 1 and a 32.7% average improvement in session 2. With only 3 target subjects, the reported aggregate MAE values could be driven by one subject or by a small number of trials. The test-split description is also ambiguous: the text says 26 participants were used in training and 3 were target subjects, but then says the dataset was partitioned 80/20 into training and validation. The authors should clarify whether target subjects' data were held out entirely, report per-subject results, and provide paired significance tests (or effect sizes with confidence intervals) for the raw-versus-normalized comparison.
  3. [Section 3.1 and Section 4.1] The assumption that per-speaker z-score normalization of F0 is the brain's speaker normalization is not tested. If the auditory system normalizes using a different transform (e.g., log-F0, percentile ranks, or an estimated speaker range), then the 'speaker-normalized' target is not the hypothesized neural code, and the decoding comparison does not test relative-pitch encoding. The authors should either include alternative normalization schemes and show the decoding advantage is specific to their chosen transform, or benchmark their target against an independent neural-encoding model. This is a correctness-risk issue rather than a circularity, but it is load-bearing for the interpretation in Section 4.1.
minor comments (5)
  1. [Introduction] There are typos and wording issues: 'poorly undared' should be 'poorly understood'; 'as showed' should be 'as shown'; 'V ocoder' should be 'vocoder'.
  2. [Section 3.1] The notation is inconsistent: the model input is defined as X in R^{C×T}, but the preprocessing text says data were standardized into matrices of dimensions (500, 60). Please clarify which dimension is time and which is channels, and note that the 64-electrode cap results in 60 channels after some processing.
  3. [Section 3.1] The session description is ambiguous: 'each speaker producing 20 monosyllabic sounds across all four tones (80 tokens per speaker)' could mean 20 syllables per tone or 20 syllables total; please state the design explicitly.
  4. [Section 3.2] The paper does not mention ethics approval, informed consent, or data/code availability. For a human EEG study, the ethics statement should be included; the dataset and code availability also need to be stated.
  5. [Section 4.1 and Figure 2] Figure 2 shows example reconstructions, but there is no quantitative description of how representative these examples are; consider adding a panel with error bounds or a side-by-side failure case.

Circularity Check

1 steps flagged · score 6.0 of 10

Reported normalized-pitch decoding advantage is partly forced by rescaling normalized predictions with oracle per-speaker mean and variance during metric calculation.

  1. fitted input called prediction [Section 3.1 (Dataset and preprocessing) and Section 4.1 (Main Results, Eq. 2)]
    "applied z-score normalization to all samples for each speaker. ... To address the significant scale differences between speaker-normalized pitch contours and raw pitch contours, which complicate direct comparisons of evaluation metrics, speaker-normalized pitch contours are rescaled to the original raw pitch contours’ scale using the speaker’s mean and variance during metric calculation. However, the training objective remains focused on speaker-normalized pitch contours."

    The normalized target is z = (F0 - μ_s)/σ_s with ground-truth per-speaker μ_s and σ_s. At evaluation, the model's normalized prediction ŷ is inverted to ŷ_raw = σ_s·ŷ + μ_s using the same speaker statistics. A zero-information model that always outputs ŷ = 0 therefore becomes a per-trial prediction of that speaker's mean F0, with MAE = E|F0 - μ_s|, while the raw-condition trivial baseline (global mean) has MAE = E|F0 - μ_g|. The gap between these baselines is the between-speaker F0 spread, which exists even if EEG carries no pitch information. The paper reports a large normalized advantage only in the six-speaker session (Table 2: Proposed raw 48.9 vs norm 30.5) and no advantage in the single-speaker session (Table 1: 18.4 vs 18.4), exactly the pattern this rescaling artifact produces.

full rationale

The identified step is the load-bearing comparison behind the paper's central claim. The z-score target and the metric-time rescaling use the same ground-truth per-speaker mean and variance, so the normalized condition is given oracle speaker statistics for free at evaluation, while the raw condition must estimate them from data. A constant-output model would produce per-speaker mean F0 predictions after inversion and would beat the raw global-mean baseline by the between-speaker F0 spread, without any EEG information. No trivial speaker-mean baseline is reported, and the single-speaker session shows no advantage, which is the signature of this artifact. This is not full definitional circularity in the network's training loss, but the headline normalized-pitch advantage is partially an evaluation-scheme construction. No problematic self-citation chain or externally forced uniqueness theorem is present; the model's raw-pitch decoding is independently meaningful, so the circularity is partial rather than total.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

No new physical entities are postulated. The central claim rests on experimental assumptions about the validity of F0 extraction, the interpretation of decoding accuracy as neural representation, and the choice of z-score normalization as the brain's normalization. These are not mathematical axioms but empirical premises that need statistical support.

free parameters (2)
  • Per-speaker z-score mean and standard deviation for F0 = Computed from each speaker's pitch contours; exact values not reported
    Defines the speaker-normalized target and rescaling; the choice of z-score transform is assumed to match the brain's normalization.
  • Training hyperparameters (learning rate, dropout, epochs, batch size, scheduler) = 5e-4, 0.2, 150, 32, cycle 20
    Chosen by hand; affect the reported accuracy but not the scientific interpretation directly.
assumptions (4)
  • domain assumption The F0 contour extracted with PyWorld is a valid ground truth for the perceived pitch of the stimulus.
    All targets and metrics are built on this extraction; errors in F0 would propagate into the comparison.
  • domain assumption Higher decoding accuracy for a target implies the brain encodes information in that target's format.
    The paper's inference from better normalized-pitch decoding to neural encoding of relative pitch assumes decoding performance reflects neural representation rather than target statistics.
  • ad hoc to paper Per-speaker z-score normalization matches the brain's speaker normalization.
    Z-scoring with mean and standard deviation is one possible normalization; no independent evidence shows the brain uses this exact transform.
  • domain assumption The preprocessed EEG contains enough stimulus-related information for single-trial pitch decoding.
    The premise of the decoding approach; no shuffled-label or chance-level control is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Decoding Speaker-Normalized Pitch from EEG for Mandarin Perception." pith.science (2026). https://pith.science/paper/TJBMIY6Y

@misc{pith2026250519626,
  author       = {Pith},
  title        = {Pith review of: Decoding Speaker-Normalized Pitch from EEG for Mandarin Perception},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TJBMIY6Y}},
  note         = {Machine review of arXiv:2505.19626}
}
read the original abstract

The same speech content produced by different speakers exhibits significant differences in pitch contour, yet listeners' semantic perception remains unaffected. This phenomenon may stem from the brain's perception of pitch contours being independent of individual speakers' pitch ranges. In this work, we recorded electroencephalogram (EEG) while participants listened to Mandarin monosyllables with varying tones, phonemes, and speakers. The CE-ViViT model is proposed to decode raw or speaker-normalized pitch contours directly from EEG. Experimental results demonstrate that the proposed model can decode pitch contours with modest errors, achieving performance comparable to state-of-the-art EEG regression methods. Moreover, speaker-normalized pitch contours were decoded more accurately, supporting the neural encoding of relative pitch.

Figures

Figures reproduced from arXiv: 2505.19626 by the authors.

Figure 1
Figure 1. The overall architecture for CE-ViViT model: (a) The dual parallel design of the temporal and channel domains in Convolu￾tional Embedding (CE). (b) The details of the Temporal, Dilation, and Channel ConvBlocks. (c) The feature fusion method of DANE. (d) Illustration of the Pitch Head. 2. Proposed Method 2.1. Overview The overall structure of the proposed CE-ViViT is illustrated in [PITH_FULL_IMAGE:figures/full_fig_… view at source ↗
Figure 2
Figure 2. Some examples of EEG-guided pitch reconstruction on session 1. First, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 30 canonical work pages

  1. [7]

    Perception of pitch location within a speaker’s f0 range,

    D. N. Honorof and D. H. Whalen, “Perception of pitch location within a speaker’s f0 range,”The Journal of the Acoustical Society of America, vol. 117, no. 4, pp. 2193–2200, 2005

    work page 2005

  2. [1]

    Decoding Speaker-Normalized Pitch from EEG for Mandarin Perception

    Introduction The fundamental frequency (F0), often referred to as pitch, is an essential and critical acoustic feature in speech signals, rep- resenting the lowest frequency component of a sound. Humans precisely control the pitch of their voices to encode linguistic meaning [1]. In non-tonal languages like English, pitch varia- tions convey intonational ...

    work page Pith review arXiv 2025

  3. [2]

    Overview The overall structure of the proposed CE-ViViT is illustrated in Figure 1, which consists of three main blocks: Convolutional Embedding, Feature Encoder, and Pitch Head

    Proposed Method 2.1. Overview The overall structure of the proposed CE-ViViT is illustrated in Figure 1, which consists of three main blocks: Convolutional Embedding, Feature Encoder, and Pitch Head. Raw EEG seg- mentsX∈R C×T serve directly as input, whereCrepresents the number of EEG channels andTdenotes the number of sam- pling points in the signal. The...

    work page

  4. [3]

    Dataset and preprocessing The stimuli comprised two monosyllabic sessions, each con- taining 480 Mandarin monosyllabic words selected from the Tone Perfect corpus 1 [25]

    Experiments 3.1. Dataset and preprocessing The stimuli comprised two monosyllabic sessions, each con- taining 480 Mandarin monosyllabic words selected from the Tone Perfect corpus 1 [25]. The sessions are described as fol- lows: •Session 1:Included 480 unique monosyllabic tokens pro- duced by a single native Mandarin male speaker, covering all four lexica...

    work page

  5. [4]

    Main Results As shown in Figure 2, these are several examples of the Man- darin pitch contours decoded from EEG by our proposed model on session 1

    Results and Analysis 4.1. Main Results As shown in Figure 2, these are several examples of the Man- darin pitch contours decoded from EEG by our proposed model on session 1. For performance comparison, we consider VLAAI [16], WaveNet [29], and HappyQuokka [17] as contrastive sys- tems, which are representative SOTA EEG-based speech regres- sion models. To...

    work page

  6. [5]

    We recorded a Mandarin auditory EEG dataset and proposed a novel auditory EEG decoding model, CE-ViViT

    Conclusion In this paper, we demonstrate that the brain’s perception of Mandarin pitch contours is independent of individual speakers’ pitch ranges. We recorded a Mandarin auditory EEG dataset and proposed a novel auditory EEG decoding model, CE-ViViT. Results on session 1 and session 2 show that the proposed model achieves performance comparable to state...

    work page

  7. [6]

    Cooperative cortical network for categorical processing of chinese lexical tone,

    X. Si, W. Zhou, and B. Hong, “Cooperative cortical network for categorical processing of chinese lexical tone,”Proceedings of the National Academy of Sciences, vol. 114, no. 46, pp. 12 303– 12 308, 2017

    work page 2017

  8. [8]

    A system of tone letters,

    Y .-R. Chao, “A system of tone letters,”Le ma ˆıtre phon ´etique, 1930

    work page 1930

Show all 35 references
  1. [9]

    Improved contextualized speech representations for tonal analysis,

    J. Yuan, X. Cai, and K. Church, “Improved contextualized speech representations for tonal analysis,” inProceedings of Interspeech, 2023, pp. 4513–4517

    2023

  2. [10]

    Pitch dynamism in female and male speech,

    C. Henton, “Pitch dynamism in female and male speech,”Lan- guage & Communication, vol. 15, no. 1, pp. 43–61, 1995

    1995

  3. [11]

    Decoding imagined musical pitch from human scalp electroencephalograms,

    M. Chung, T. Kim, E. Jeong, C. K. Chung, J. S. Kim, O.-S. Kwon, and S.-P. Kim, “Decoding imagined musical pitch from human scalp electroencephalograms,”IEEE Transactions on Neural Sys- tems and Rehabilitation Engineering, vol. 31, pp. 2154–2163, 2023

    2023

  4. [12]

    Dbpnet: Dual-branch parallel network with temporal-frequency fusion for auditory attention detection,

    Q. Ni, H. Zhang, C. Fan, S. Pei, C. Zhou, and Z. Lv, “Dbpnet: Dual-branch parallel network with temporal-frequency fusion for auditory attention detection,” inProceedings of the International Joint Conference on Artificial Intelligence (IJCAI 2024), 2024

    2024

  5. [13]

    Intonational speech prosody encoding in the human auditory cortex,

    C. Tang, L. Hamilton, and E. Chang, “Intonational speech prosody encoding in the human auditory cortex,”Science, vol. 357, no. 6353, pp. 797–801, 2017

    2017

  6. [14]

    Human cortical encoding of pitch in tonal and non-tonal languages,

    Y . Li, C. Tang, J. Lu, J. Wu, and E. F. Chang, “Human cortical encoding of pitch in tonal and non-tonal languages,”Nature com- munications, vol. 12, no. 1, p. 1161, 2021

    2021

  7. [15]

    Decoding and synthesizing tonal language speech from brain activity,

    Y . Liu, Z. Zhao, M. Xu, H. Yu, Y . Zhu, J. Zhang, L. Bu, X. Zhang, J. Lu, Y . Liet al., “Decoding and synthesizing tonal language speech from brain activity,”Science Advances, vol. 9, no. 23, p. eadh0478, 2023

    2023

  8. [16]

    Synthesiz- ing speech from ecog with a combination of transformer-based encoder and neural vocoder,

    K. Shigemi, S. Komeiji, T. Mitsuhashi, Y . Iimura, H. Suzuki, H. Sugano, K. Shinoda, K. Yatabe, and T. Tanaka, “Synthesiz- ing speech from ecog with a combination of transformer-based encoder and neural vocoder,” inICASSP 2023-2023 IEEE Inter- national Conference on Acoustics,...

    2023

  9. [17]

    Decoding speech perception from non-invasive brain record- ings,

    A. D ´efossez, C. Caucheteux, J. Rapin, O. Kabeli, and J.-R. King, “Decoding speech perception from non-invasive brain record- ings,”Nature Machine Intelligence, vol. 5, no. 10, pp. 1097–1107, 2023

    2023

  10. [18]

    Convconcatnet: a deep convolutional neural network to reconstruct mel spectrogram from the eeg,

    X. Xu, B. Wang, Y . Yan, H. Zhu, Z. Zhang, X. Wu, and J. Chen, “Convconcatnet: a deep convolutional neural network to reconstruct mel spectrogram from the eeg,”arXiv preprint arXiv:2401.04965, 2024

    2024 arXiv

  11. [19]

    Modeling the relationship between acoustic stimulus and eeg with a dilated convolutional neural network,

    B. Accou, M. J. Monesi, J. Montoya, T. Francartet al., “Modeling the relationship between acoustic stimulus and eeg with a dilated convolutional neural network,” in2020 28th European Signal Pro- cessing Conference (EUSIPCO). IEEE, 2021, pp. 1175–1179

    2021

  12. [20]

    Naturalistic music decoding from eeg data via latent diffusion models,

    E. Postolache, N. Polouliakh, H. Kitano, A. Connelly, E. Rodol `a, L. Cosmo, and T. Akama, “Naturalistic music decoding from eeg data via latent diffusion models,”arXiv preprint arXiv:2405.09062, 2024

    2024 arXiv

  13. [21]

    Cross-subject classification of spoken mandarin vowels and tones with eeg sig- nals: A study of end-to-end cnn with fine-tuning,

    X. Wang, M. Li, H. Li, S. H. Pun, and F. Chen, “Cross-subject classification of spoken mandarin vowels and tones with eeg sig- nals: A study of end-to-end cnn with fine-tuning,” in2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APS...

    2023

  14. [22]

    The detailed architecture of DANE is illustrated in Figure 1(c)

    from the Doubly Fused ViT (DFvT), applying separate at- tention mechanisms to both temporal and channel outputs to generate the embeddingV. The detailed architecture of DANE is illustrated in Figure 1(c). The Temporal ConvBlock, Dilation ConvBlock, and Chan- nel ConvBlock shar...

  15. [23]

    De- coding of the speech envelope from eeg using the vlaai deep neu- ral network,

    B. Accou, J. Vanthornhout, H. V . hamme, and T. Francart, “De- coding of the speech envelope from eeg using the vlaai deep neu- ral network,”Scientific Reports, vol. 13, no. 1, p. 812, 2023

    2023

  16. [24]

    Happyquokka sys- tem for icassp 2023 auditory eeg challenge,

    Z. Piao, M. Kim, H. Yoon, and H.-G. Kang, “Happyquokka sys- tem for icassp 2023 auditory eeg challenge,” inICASSP 2023- 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023, pp. 1–2

    2023

  17. [25]

    Magnetoencephalography—theory, instru- mentation, and applications to noninvasive studies of the working human brain,

    M. H ¨am¨al¨ainen, R. Hari, R. J. Ilmoniemi, J. Knuutila, and O. V . Lounasmaa, “Magnetoencephalography—theory, instru- mentation, and applications to noninvasive studies of the working human brain,”Reviews of modern Physics, vol. 65, no. 2, p. 413, 1993

    1993

  18. [26]

    Encoding and decoding neuronal dynamics: Methodological framework to uncover the algorithms of cognition,

    J.-R. King, A. Gramfortet al., “Encoding and decoding neuronal dynamics: Methodological framework to uncover the algorithms of cognition,” 2018

    2018

  19. [27]

    Scconv: Spatial and channel reconstruc- tion convolution for feature redundancy,

    J. Li, Y . Wen, and L. He, “Scconv: Spatial and channel reconstruc- tion convolution for feature redundancy,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2023, pp. 6153–6162

    2023

  20. [28]

    Doubly-fused vit: Fuse infor- mation from vision transformer doubly with local representation,

    L. Gao, D. Nie, B. Li, and X. Ren, “Doubly-fused vit: Fuse infor- mation from vision transformer doubly with local representation,” inEuropean Conference on Computer Vision. Springer, 2022, pp. 744–761

    2022

  21. [29]

    Vivit: A video vision transformer,

    A. Arnab, M. Dehghani, G. Heigold, C. Sun, M. Lu ˇci´c, and C. Schmid, “Vivit: A video vision transformer,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 6836–6846

    2021

  22. [30]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,”Advances in Neural In- formation Processing Systems, 2017

    2017

  23. [31]

    Tone perfect: Multimodal database for mandarin chinese,

    C. Ryu, the Mandarin Tone Perception & Production Team, and M. S. U. Libraries, “Tone perfect: Multimodal database for mandarin chinese,” Michigan State University Libraries, 2022, accessed: 1 January 2022. [Online]. Available: https: //tone.lib.msu.edu/

    2022

  24. [32]

    Eeglab: an open source toolbox for analysis of single-trial eeg dynamics including independent component analysis,

    A. Delorme and S. Makeig, “Eeglab: an open source toolbox for analysis of single-trial eeg dynamics including independent component analysis,”Journal of neuroscience methods, vol. 134, no. 1, pp. 9–21, 2004

    2004

  25. [33]

    Speech synthesis using eeg,

    G. Krishna, C. Tran, Y . Han, M. Carnahan, and A. H. Tewfik, “Speech synthesis using eeg,” inICASSP 2020-2020 IEEE Inter- national Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2020, pp. 1235–1238

    2020

  26. [34]

    Comparative performance study of several pitch detection algo- rithms,

    M. J. Cheng, L. Rabiner, A. Rosenberg, and C. McGonegal, “Comparative performance study of several pitch detection algo- rithms,”The Journal of the Acoustical Society of America, vol. 58, no. S1, pp. S61–S62, 1975

    1975

  27. [35]

    Decoding auditory eeg responses using an adapted wavenet,

    B. Van Dyck, L. Yang, and M. M. Van Hulle, “Decoding auditory eeg responses using an adapted wavenet,” inICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023, pp. 1–2

    2023

Pith tools

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