Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

From Thought to Action: How a Hierarchy of Neural Dynamics Supports Language Production

T0 review · 3 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read During typing, the brain activates context, word, syllable, and letter representations in a top-down sequence, each sustained and superposed, coordinated by hierarchical dynamic neural codes.

desk verdict Strong empirical work on dynamic neural codes in typing, but the claimed top-down ordering rests on half-times that are likely contaminated by sustained context/word activity from the reading phase. read the letter →

arxiv 2502.07429 v2 pith:B6O2X3PL submitted 2025-02-11 q-bio.NC

Mingfang Zhang , Jarod Lévy , Stéphane d'Ascoli , Jérémy Rapin , F.-Xavier Alario , Pierre Bourdillon , Svetlana Pinet , Jean-Rémi King This is my paper · ORCID
classification q-bio.NC
keywords languageproductionneuraldecodingMEGEEGtypingdynamiccodetemporalgeneralizationhierarchicalrepresentation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that language production is not a single burst of motor commands but a cascade: before each typed word, the brain first builds a context representation, then a word representation, then syllables, then letters. Using MEG and EEG recorded while 35 skilled typists typed memorized sentences, the authors linearly decode these four levels from brain signals. They find the levels rise and fall in order, each remains decodable for seconds—far longer than the keystroke it drives—so representations of several upcoming and past items overlap in time. The overlap is resolved by a 'dynamic neural code': each representation continuously moves to a different neural subspace, with higher levels moving more slowly. If correct, this gives a computational account of how thought becomes action.

What carries the argument

Temporal generalization analysis is the load-bearing tool: a decoder trained at time t is tested at all other times; a diagonal generalization matrix means the code is dynamic (the representation moves), a square matrix means it is static. The paper combines this with 'half-times'—the moment decoding reaches 50% of its peak—to order the rise of representations, and with shifted temporal generalization to test whether a decoder for one item can read out its neighbors.

What would settle it

Compute half-times after subtracting the average decoding level measured during the wait period before typing a sentence. If the context-to-letter ordering disappears or inverts, the reported top-down sequence reflects sustained memory rather than a production cascade.

Watch

Extended reading notes

Core claim

The central discovery is a hierarchy of dynamic neural codes that schedules language production. Decoding context (from GPT-2), word (Spacy), syllable (FastText), and letter (one-hot) embeddings from MEG signals yields half-times that order by level: context first, then word, syllable, letter, with a Spearman correlation R=-0.77 across subjects; the same ordering appears in EEG. Separate temporal-generalization analyses show each representation is stable in availability but displaced in neural space over time, and the generalization width scales with level—letters generalize for ~280 ms, contexts for ~780 ms. Successive letters up to five positions apart can be decoded from the same instant, and the brain resolves this superposition by keeping each item in its own moving subspace.

Load-bearing premise

The half-time ordering assumes that the moment a representation reaches half its peak reflects the start of a new representation formed for the current word; if context and word representations are carried over from reading or from the previous word, their early half-times would be an artifact.

Editorial extensions

If this is right

  • If the cascade is real, typing offers an artifact-free window onto the neural staging of language production that speech artifacts obscure.
  • The same top-down ordering should appear in other production modalities such as speaking, writing, and signing, with comparable time constants.
  • Dynamic coding explains how many overlapping serial-order representations can coexist without mutual interference, a general problem beyond language.
  • The level-dependent speed of neural dynamics predicts that disrupting a higher-level representation should delay lower-level decoding by predictable amounts.
  • The reported decodability of upcoming letters and words supports brain-to-text interfaces that read out intended sequences before execution.

Reading between the lines

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

  • Editorial inference: the speed gradient (letters fastest, contexts slowest) may reflect not just linguistic abstraction but also the temporal granularity of the actions each level controls; comparing typing with handwriting or speech could dissociate these factors.
  • Editorial inference: the moving-subspace code could be implemented by traveling waves or oscillatory phase gradients, a prediction testable with intracranial recordings or high-density MEG.
  • Editorial inference: the half-time comparisons may partly reflect memory maintenance from the read phase; an explicit control that varies the retention interval between reading and typing would sharpen the production-specific claim.
  • Editorial inference: if the same dynamic hierarchy appears in speech comprehension (slower phrases, faster phonemes), it would suggest a canonical brain strategy for representing nested sequential structure in both perception and production.
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 / 6 minor

Summary. The manuscript investigates the neural dynamics of language production using MEG and EEG recordings from 35 skilled typists performing a read-wait-type task. The authors linearly decode four hierarchical levels of linguistic representations—context (GPT-2), word (Spacy), syllable (FastText), and letter (one-hot)—and report that before each typed word, these representations rise sequentially from context to letter, while overlapping in time. They further report evidence for a hierarchy of dynamic neural codes, with higher-level representations showing slower temporal dynamics than lower-level ones, and for superposition of successive representations during typing. The central claim is that language production is supported by a top-down cascade of neural representations that are maintained over long periods and coordinated by a hierarchy of dynamic codes.

Significance. If the central claim holds, this would be an important non-invasive, whole-brain demonstration of the hypothesized hierarchical organization of language production, extending prior work from picture naming and single-word tasks to natural multi-word sentence typing. The study has notable strengths: the decoding targets come from external pretrained models (GPT-2, Spacy, FastText) rather than being fit to the neural data, the train/test splits are by sentence to reduce leakage, and the main results are replicated in an independent EEG dataset. The paper also clearly states its limitations regarding anatomical localization and modality generalizability. However, the interpretation of the half-time ordering as a production-triggered sequential rise is threatened by a baseline-carryover confound that is not controlled, and the superposition claims lack a serial-position control. These issues are load-bearing for the main message, but they are addressable with additional analyses, so the contribution is potentially valuable if those controls confirm the original conclusions.

major comments (3)
  1. [Section 4.3 ('Half-times') and Figure 2A-B] The half-time measure is computed within a fixed [-1, 0] s window without any baseline correction, but the task is read-wait-type and Figure S3 itself shows significant context and word decoding from -1.5 s, i.e. before the analysis window begins. If the decoding score at the start of the window already exceeds 50% of the within-window maximum, the half-time is pinned to the window edge, so the observed ordering (context before word before syllable before letter) and the Spearman correlation R=-0.77, p<10^-5 could reflect sustained activity carried over from the reading phase rather than a production-triggered sequential rise. The paper does not report the distribution of half-times relative to the window boundaries, nor any control such as analyzing the first word after the 1.5 s delay in isolation, nor a baseline-subtracted version of the decoding curves. This is load-bearing for the central claim of a top-down cascade, so the authors should provide these controls or explicitly qualify the half-time interpretation.
  2. [Section 2 ('Simultaneous representations of multiple keys during word production') and Figure 2C-D] The shifted-time decoding analysis shows that successive letters can be decoded from the same time sample, but there is no serial-position control. In this read-wait-type task, a single MEG/EEG time sample contains motor planning and execution signals for the current key, proprioceptive and visual feedback from the previous key, and possibly pre-activation of upcoming keys; these non-orthographic signals are structured by serial position and could produce overlapping decodability of neighboring letters without any simultaneous representation of multiple letter identities. The paper should add controls such as (a) decoding shuffled letter orders within the same word, (b) comparing with a null model that only represents the currently executed key, or (c) using natural variation in inter-key intervals to dissociate time from serial position.
  3. [Section 2 ('Letters are represented with a dynamic neural code' and 'A hierarchy of dynamical codes') and Figures 3B…] The diagonal temporal-generalization pattern is interpreted as evidence for moving neural subspaces, but a diagonal pattern can also arise from slow non-stationarities in the signal (e.g., evoked response ramps, attentional shifts, or motor preparation) that are not specific to a representational code. The paper does not include a control such as testing whether a decoder trained at one time generalizes to an independent condition, or comparing the observed matrices with surrogate data obtained by shuffling labels across time. Without such a control, the speed-hierarchy correlation (R=0.74, p<10^-5) is difficult to interpret as a property of the neural code itself rather than a reflection of the time-varying signal envelope.
minor comments (6)
  1. [Section 2 (Decomposing the hierarchy)] The text reports 'R=0.17±0.01%' and 'R=0.13±0.01%' for word and context decoding; the percent sign is inappropriate for a Pearson correlation and should be removed.
  2. [Section 2 (Sequential deactivations)] The phrase 'afterword offset' should be 'after word offset'.
  3. [Section 2 (Simultaneous representations)] The sentence 'he last three keys can be decoded after word offset' contains a typo and should read 'the last three keys'.
  4. [Section 3 (Discussion)] The sentence ends with 'remains to be resolved..'—the double period should be a single period.
  5. [Section 4.3 (Significance testing)] The text says 'False Detection Rate (FDR)'; the standard term is 'False Discovery Rate', and this should be corrected.
  6. [Figure 2B caption] The phrase 'skip-level comparisons' is not defined; please explain which pairwise comparisons are shown and why only some are displayed, or refer to the supplementary material for the full set.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the hierarchy claims rest on external embeddings and held-out neural decoding, not on fitted parameters or self-citation chains.

full rationale

The paper's central claims are derived from linear decoding of external linguistic embeddings (GPT-2, Spacy, FastText, one-hot) from MEG/EEG signals, with within-subject group k-fold cross-validation split by sentence. The half-time ordering is an empirical summary of held-out decoding curves, not a parameter fitted to the claimed hierarchy; the Spearman correlations between half-times and linguistic level are post-hoc summaries, not construction. The temporal generalization method is cited to King and Dehaene (2014), a standard published method, and the dynamic-code interpretation is supported by the structure of cross-decoding matrices rather than by an imported uniqueness theorem. Citations to the authors' own prior work (e.g., Caucheteux and King 2022 for the GPT-2 layer choice, the companion paper Levy et al. 2025 for a decoding pipeline, and shared Fig. 1A/D) are not load-bearing for the hierarchy claim. The reader's concern that half-times computed in [-1,0] may be biased by representations carried over from the reading phase (as suggested by significant decoding already at -1.5s in Figure S3) is a potential baseline/validity confound, but it is not a circular reduction: the paper does not define the hierarchy in terms of the half-time measure, nor fit the ordering. The derivation chain is therefore self-contained with respect to the circularity patterns considered here.

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

The paper relies on standard operational assumptions in cognitive neuroscience: that linear decodability from M/EEG indicates neural representation, that pre-trained embedding spaces approximate linguistic content, that typing is a valid model of language production, and that temporal generalization diagonals indicate dynamic codes. It introduces no new free parameters or entities.

assumptions (4)
  • domain assumption The embedding spaces of pre-trained language models (GPT-2 layer 8, Spacy, FastText) capture the neural representational content of context, word, and syllable levels.
    The paper uses these models as target features (§4.2) and assumes linear decodability from brain signals implies the brain uses these codes.
  • domain assumption Linear decodability defines neural representation (DiCarlo and Cox, 2007).
    All conclusions about representation depend on this operational definition.
  • domain assumption Typing engages the same language production machinery as speaking.
    The authors acknowledge typing is not natural for evolution (Discussion), and letter representations may be keyboard-specific motor actions.
  • domain assumption A diagonal temporal generalization matrix indicates a dynamic neural code.
    The interpretation of TG patterns as evidence for changing subspaces assumes no other time-varying confounds (evoked responses, attention) produce diagonality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Thought to Action: How a Hierarchy of Neural Dynamics Supports Language Production." pith.science (2026). https://pith.science/paper/B6O2X3PL

@misc{pith2026250207429,
  author       = {Pith},
  title        = {Pith review of: From Thought to Action: How a Hierarchy of Neural Dynamics Supports Language Production},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B6O2X3PL}},
  note         = {Machine review of arXiv:2502.07429}
}
read the original abstract

Humans effortlessly communicate their thoughts through intricate sequences of motor actions. Yet, the neural processes that coordinate language production remain largely unknown, in part because speech artifacts limit the use of neuroimaging. To elucidate the unfolding of language production in the brain, we investigate with magnetoencephalography (MEG) and electroencephalography (EEG) the neurophysiological activity of 35 skilled typists, while they typed sentences on a keyboard. This approach confirms the hierarchical predictions of linguistic theories: the neural activity preceding the production of each word is marked by the sequential rise and fall of context-, word-, syllable-, and letter-level representations. Remarkably, each of these neural representations is maintained over long time periods within each level of the language hierarchy. This phenomenon results in a superposition of successive representations that is supported by a hierarchy of dynamic neural codes. Overall, these findings provide a precise computational breakdown of the neural dynamics that coordinate the production of language in the human brain.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception

    cs.CV 2025-08 unverdicted novelty 5.0 of 10

    WorldWeaver reduces temporal drift in long-horizon video generation by jointly modeling RGB and depth perceptual conditions with segmented noise scheduling.

  2. Foundation Models for Cross-Domain EEG Analysis Application: A Survey

    cs.HC 2025-08 conditional novelty 4.0 of 10

    A survey that organizes EEG foundation-model research into five output-modality categories: native EEG, text, vision, audio, and multimodal fusion, with a claim to be the first such comprehensive taxonomy.

Reference graph

Works this paper leans on

57 extracted references · 37 canonical work pages · cited by 2 Pith papers

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Correcting MEG Artifacts Caused by Overt Speech

    Omid Abbasi, Nadine Steingräber, and Joachim Gross. Correcting MEG Artifacts Caused by Overt Speech . Frontiers in Neuroscience, 15: 0 682419, June 2021. ISSN 1662-453X. doi:10.3389/fnins.2021.682419. https://www.frontiersin.org/articles/10.3389/fnins.2021.682419/full

    work page arXiv 2021

  3. [3]

    Averbeck, Matthew V Chafee, David A

    Bruno B. Averbeck, Matthew V Chafee, David A. Crowe, and Apostolos P Georgopoulos. Parallel processing of serial movements in prefrontal cortex. Proceedings of the National Academy of Sciences of the United States of America, 99 0 (20): 0 13172--13177, 2002. ISSN 00278424. doi:10.1073/pnas.162485599. ISBN: 0027-8424

    work page doi:10.1073/pnas.162485599 2002

  4. [4]

    Language switching decomposed through MEG and evidence from bimodal bilinguals

    Esti Blanco-Elorrieta, Karen Emmorey, and Liina Pylkkänen. Language switching decomposed through MEG and evidence from bimodal bilinguals. Proceedings of the National Academy of Sciences, 115 0 (39): 0 9708--9713, September 2018. doi:10.1073/pnas.1809779115. https://www.pnas.org/doi/10.1073/pnas.1809779115. Publisher: Proceedings of the National Academy o...

    work page doi:10.1073/pnas.1809779115 2018

  5. [5]

    Language production: Grammatical encoding

    Kathryn Bock and Willem Levelt. Language production: Grammatical encoding. In M.A. Gemsbacher, editor, Handbook of psycholinguistics, pages 945--984. Academic Press, San Diego, 1994

    work page 1994

  6. [6]

    Enriching Word Vectors with Subword Information , June 2017

    Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. Enriching Word Vectors with Subword Information , June 2017. http://arxiv.org/abs/1607.04606. arXiv:1607.04606 [cs]

    arXiv 2017

  7. [7]

    Bourguignon

    Nicolas J. Bourguignon. A rostro-caudal axis for language in the frontal lobe: the role of executive control in speech production. Neuroscience and Biobehavioral Reviews, 47: 0 431--444, November 2014. ISSN 1873-7528. doi:10.1016/j.neubiorev.2014.09.008

    work page doi:10.1016/j.neubiorev.2014.09.008 2014

  8. [8]

    The Time Course of Language Production as Revealed by Pattern Classification of MEG Sensor Data

    Francesca Carota, Jan-Mathijs Schoffelen, Robert Oostenveld, and Peter Indefrey. The Time Course of Language Production as Revealed by Pattern Classification of MEG Sensor Data . The Journal of Neuroscience, 42 0 (29): 0 5745--5754, July 2022. ISSN 0270-6474, 1529-2401. doi:10.1523/JNEUROSCI.1923-21.2022. https://www.jneurosci.org/lookup/doi/10.1523/JNEUR...

    work page doi:10.1523/jneurosci.1923-21.2022 2022

Show all 57 references
  1. [9]

    Language processing in brains and deep neural networks: computational convergence and its limits, January 2021

    Charlotte Caucheteux and Jean-Rémi King. Language processing in brains and deep neural networks: computational convergence and its limits, January 2021. https://www.biorxiv.org/content/10.1101/2020.07.03.186288v2. Pages: 2020.07.03.186288 Section: New Results

    2021 doi

  2. [10]

    Brains and algorithms partially converge in natural language processing

    Charlotte Caucheteux and Jean-Rémi King. Brains and algorithms partially converge in natural language processing. Communications Biology, 5 0 (1): 0 1--10, February 2022. ISSN 2399-3642. doi:10.1038/s42003-022-03036-1. https://www.nature.com/articles/s42003-022-03036-1. Publis...

    2022 doi

  3. [11]

    Claire H. C. Chang, Samuel A. Nastase, and Uri Hasson. Information flow across the cortical timescale hierarchy during narrative construction. Proceedings of the National Academy of Sciences, 119 0 (51): 0 e2209307119, December 2022. doi:10.1073/pnas.2209307119. https://www.pn...

    2022 doi

  4. [12]

    Aspects of the theory of syntax

    Noam Chomsky. Aspects of the theory of syntax. Aspects of the theory of syntax. M.I.T. Press, Oxford, England, 1965

    1965

  5. [13]

    Churchland, John P

    Mark M. Churchland, John P. Cunningham, Matthew T. Kaufman, Justin D. Foster, Paul Nuyujukian, Stephen I. Ryu, and Krishna V. Shenoy. Neural population dynamics during reaching. Nature, 487 0 (7405): 0 51--56, July 2012. ISSN 1476-4687. doi:10.1038/nature11129. https://www.nat...

    2012 doi

  6. [14]

    Lauren Cloutman, Leila Gingis, Melissa Newhart, Cameron Davis, Jennifer Heidler-Gary, Jennifer Crinion, and Argye E. Hillis. A Neural Network Critical for Spelling . Annals of neurology, 66 0 (2): 0 249--253, August 2009. ISSN 0364-5134. doi:10.1002/ana.21693. https://www.ncbi...

    2009 doi

  7. [15]

    Crowe, Wilbert Zarco, Ramon Bartolo, and Hugo Merchant

    David A. Crowe, Wilbert Zarco, Ramon Bartolo, and Hugo Merchant. Dynamic Representation of the Temporal and Sequential Structure of Rhythmic Movements in the Primate Medial Premotor Cortex . Journal of Neuroscience, 34 0 (36): 0 11972--11983, September 2014. ISSN 0270-6474, 15...

    2014 doi

  8. [16]

    Dell, Lisa K

    Gary S. Dell, Lisa K. Burger, and William R. Svec. Language production and serial order: A functional analysis and a model. Psychological Review, 104 0 (1): 0 123--147, January 1997. ISSN 1939-1471, 0033-295X. doi:10.1037/0033-295X.104.1.123. https://doi.apa.org/doi/10.1037/00...

    1997 doi

  9. [17]

    DiCarlo and David D

    James J. DiCarlo and David D. Cox. Untangling invariant object recognition. Trends in Cognitive Sciences, 11 0 (8): 0 333--341, August 2007. ISSN 13646613. doi:10.1016/j.tics.2007.06.010. https://linkinghub.elsevier.com/retrieve/pii/S1364661307001593

    2007 doi

  10. [18]

    Grabowski

    Karen Emmorey, Sonya Mehta, and Thomas J. Grabowski. The neural correlates of sign versus word production. NeuroImage, 36 0 (1): 0 202--208, May 2007. ISSN 1053-8119. doi:10.1016/j.neuroimage.2007.02.040

    2007 doi

  11. [19]

    Shared or different: How linked are word production and comprehension? TIPA

    Amie Fairs, Raphaël Fargier, and Kristof Strijkers. Shared or different: How linked are word production and comprehension? TIPA. Travaux interdisciplinaires sur la parole et le langage, 0 (38), December 2022. ISSN 1621-0360. doi:10.4000/tipa.4879. https://journals.openedition....

    2022 doi

  12. [20]

    Ivanova, and Tamar I

    Evelina Fedorenko, Anna A. Ivanova, and Tamar I. Regev. The language network as a natural kind within the broader landscape of the human brain. Nature Reviews Neuroscience, pages 1--24, April 2024. ISSN 1471-0048. doi:10.1038/s41583-024-00802-4. https://www.nature.com/articles...

    2024 doi

  13. [21]

    Ariel Goldstein, Zaid Zada, Eliav Buchnik, Mariano Schain, Amy Price, Bobbi Aubrey, Samuel A. Nastase, Amir Feder, Dotan Emanuel, Alon Cohen, Aren Jansen, Harshvardhan Gazula, Gina Choe, Aditi Rao, Catherine Kim, Colton Casto, Lora Fanda, Werner Doyle, Daniel Friedman, Patrici...

    2022

  14. [22]

    MEG and EEG data analysis with MNE - Python

    Alexandre Gramfort. MEG and EEG data analysis with MNE - Python . Frontiers in Neuroscience, 7, 2013. ISSN 1662453X. doi:10.3389/fnins.2013.00267. http://journal.frontiersin.org/article/10.3389/fnins.2013.00267/abstract

    2013

  15. [23]

    Guenther

    Frank H. Guenther. Neural Control of Speech . The MIT Press, July 2016. ISBN 978-0-262-33698-7. doi:10.7551/mitpress/10471.001.0001. https://direct.mit.edu/books/monograph/4085/Neural-Control-of-Speech

    2016 doi

  16. [24]

    Hierarchical dynamic coding coordinates speech comprehension in the brain, April 2024

    Laura Gwilliams, Alec Marantz, David Poeppel, and Jean-Remi King. Hierarchical dynamic coding coordinates speech comprehension in the brain, April 2024. https://www.biorxiv.org/content/10.1101/2024.04.19.590280v1. Pages: 2024.04.19.590280 Section: New Results

    2024 doi

  17. [25]

    WP2TXT : A command-line toolkit to extract text content and category data from Wikipedia dump files, 2023

    Yoichiro Hasebe. WP2TXT : A command-line toolkit to extract text content and category data from Wikipedia dump files, 2023

    2023

  18. [26]

    Computational neuroanatomy of speech production

    Gregory Hickok. Computational neuroanatomy of speech production. Nature Reviews Neuroscience, 13 0 (2): 0 135--145, February 2012. ISSN 1471-0048. doi:10.1038/nrn3158. https://www.nature.com/articles/nrn3158. Publisher: Nature Publishing Group

    2012 doi

  19. [27]

    The architecture of speech production and the role of the phoneme in speech processing

    Gregory Hickok. The architecture of speech production and the role of the phoneme in speech processing. Language and cognitive processes, 29 0 (1): 0 2--20, January 2014. ISSN 0169-0965. doi:10.1080/01690965.2013.834370. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3904400/

    2014

  20. [28]

    Precision fMRI reveals that the language-selective network supports both phrase-structure building and lexical access during language production

    Jennifer Hu, Hannah Small, Hope Kean, Atsushi Takahashi, Leo Zekelman, Daniel Kleinman, Elizabeth Ryan, Alfonso Nieto-Castañón, Victor Ferreira, and Evelina Fedorenko. Precision fMRI reveals that the language-selective network supports both phrase-structure building and lexica...

    2023 doi

  21. [29]

    The Spatial and Temporal Signatures of Word Production Components : A Critical Update

    Peter Indefrey. The Spatial and Temporal Signatures of Word Production Components : A Critical Update . Frontiers in Psychology, 2, 2011. ISSN 1664-1078. doi:10.3389/fpsyg.2011.00255. http://journal.frontiersin.org/article/10.3389/fpsyg.2011.00255/abstract

    2011 arXiv

  22. [30]

    FastText .zip: Compressing text classification models, December 2016

    Armand Joulin, Edouard Grave, Piotr Bojanowski, Matthijs Douze, Hérve Jégou, and Tomas Mikolov. FastText .zip: Compressing text classification models, December 2016. http://arxiv.org/abs/1612.03651. arXiv:1612.03651 [cs]

    2016 arXiv

  23. [31]

    Khanna, William Muñoz, Young Joon Kim, Yoav Kfir, Angelique C

    Arjun R. Khanna, William Muñoz, Young Joon Kim, Yoav Kfir, Angelique C. Paulk, Mohsen Jamali, Jing Cai, Martina L. Mustroph, Irene Caprara, Richard Hardstone, Mackenna Mejdell, Domokos Meszéna, Abigail Zuckerman, Jeffrey Schweitzer, Sydney Cash, and Ziv M. Williams. Single-neu...

    2024 doi

  24. [32]

    King and S

    J-R. King and S. Dehaene. Characterizing the dynamics of mental representations: the temporal generalization method. Trends in cognitive sciences, 18 0 (4): 0 203--210, April 2014. ISSN 1364-6613. doi:10.1016/j.tics.2014.01.002. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5635958/

    2014 doi

  25. [33]

    Meyer, Anna Sadnicka, Gareth Barnes, and Neil Burgess

    Katja Kornysheva, Daniel Bush, Sofie S. Meyer, Anna Sadnicka, Gareth Barnes, and Neil Burgess. Neural Competitive Queuing of Ordinal Structure Underlies Skilled Sequential Action . Neuron, 101 0 (6): 0 1166--1180.e3, March 2019. ISSN 0896-6273. doi:10.1016/j.neuron.2019.01.018...

    2019 doi

  26. [34]

    Willem J. M. Levelt and Antje S. Meyer. Word for word: Multiple lexical access in speech production. European Journal of Cognitive Psychology, December 2000. doi:10.1080/095414400750050178. https://www.tandfonline.com/doi/abs/10.1080/095414400750050178. Publisher: Taylor & Fra...

    doi

  27. [35]

    Anumanchipalli, Abdelrahman Mohamed, Peili Chen, Laurel H

    Yuanning Li, Gopala K. Anumanchipalli, Abdelrahman Mohamed, Peili Chen, Laurel H. Carney, Junfeng Lu, Jinsong Wu, and Edward F. Chang. Dissecting neural computations in the human auditory pathway using deep neural networks for speech. Nature Neuroscience, 26 0 (12): 0 2213--22...

    2023 doi

  28. [36]

    Neural decoding of speech with semantic-based classification

    Yi Lin and Po-Jang Hsieh. Neural decoding of speech with semantic-based classification. Cortex, 154: 0 231--240, September 2022. ISSN 0010-9452. doi:10.1016/j.cortex.2022.05.018. https://www.sciencedirect.com/science/article/pii/S0010945222001666

    2022 doi

  29. [37]

    Logan and Matthew J.C

    Gordon D. Logan and Matthew J.C. Crump. Hierarchical Control of Cognitive Processes . In Psychology of Learning and Motivation , volume 54, pages 1--27. Elsevier, 2011. ISBN 978-0-12-385527-5. doi:10.1016/B978-0-12-385527-5.00001-2. https://linkinghub.elsevier.com/retrieve/pii...

    2011 doi

  30. [38]

    Brain-to- Text Decoding : A Non -invasive Approach via Typing

    Jarod Lévy, Mingfang(Lucy) Zhang, Svetlana Pinet, Jérémy Rapin, Hubert Banville, Stéphane d'Ascoli, and Jean-Rémi King. Brain-to- Text Decoding : A Non -invasive Approach via Typing . 2025

    2025

  31. [39]

    Stark, Gregory Hickok, and Julius Fridriksson

    William Matchin, Alexandra Basilakos, Dirk-Bart den Ouden, Brielle C. Stark, Gregory Hickok, and Julius Fridriksson. Functional differentiation in the language network revealed by lesion-symptom mapping. NeuroImage, 247: 0 118778, February 2022. ISSN 1053-8119. doi:10.1016/j.n...

    2022

  32. [40]

    Toward a realistic model of speech processing in the brain with self-supervised learning, March 2023

    Juliette Millet, Charlotte Caucheteux, Pierre Orhan, Yves Boubenec, Alexandre Gramfort, Ewan Dunbar, Christophe Pallier, and Jean-Remi King. Toward a realistic model of speech processing in the brain with self-supervised learning, March 2023. http://arxiv.org/abs/2206.01685. a...

    2023 arXiv

  33. [41]

    explosion/ spaCy : v3.7.2: Fixes for APIs and requirements, October 2023

    Ines Montani, Matthew Honnibal, Adriane Boyd, Sofie Van Landeghem, and Henning Peters. explosion/ spaCy : v3.7.2: Fixes for APIs and requirements, October 2023. https://zenodo.org/doi/10.5281/zenodo.1212303

    2023 doi

  34. [42]

    On the cortical dynamics of word production: a review of the MEG evidence

    Dashiel Munding, Anne-Sophie Dubarry, and F.-Xavier Alario. On the cortical dynamics of word production: a review of the MEG evidence. Language, Cognition and Neuroscience, 31 0 (4): 0 441--462, April 2016. ISSN 2327-3798. doi:10.1080/23273798.2015.1071857. https://www.tandfon...

    2016

  35. [43]

    Scikit-learn: Machine Learning in Python

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesna...

    2011

  36. [44]

    Response retrieval and motor planning during typing

    Svetlana Pinet, Anne-Sophie Dubarry, and F.-Xavier Alario. Response retrieval and motor planning during typing. Brain and Language, 159: 0 74--83, August 2016. ISSN 0093934X. doi:10.1016/j.bandl.2016.05.012. https://linkinghub.elsevier.com/retrieve/pii/S0093934X15301802

    2016 doi

  37. [45]

    Dell, and F.-Xavier Alario

    Svetlana Pinet, Gary S. Dell, and F.-Xavier Alario. Tracking Keystroke Sequences at the Cortical Level Reveals the Dynamics of Serial Order Production . Journal of Cognitive Neuroscience, 31 0 (7): 0 1030--1043, July 2019. ISSN 0898-929X. doi:10.1162/jocn_a_01401. https://doi....

    2019 doi

  38. [46]

    handwriting brain

    Samuel Planton, Mélanie Jucla, Franck-Emmanuel Roux, and Jean-François Démonet. The “handwriting brain”: A meta-analysis of neuroimaging studies of motor versus orthographic processes. Cortex, 49 0 (10): 0 2772--2787, November 2013. ISSN 0010-9452. doi:10.1016/j.cortex.2013.05...

    2013 doi

  39. [47]

    Cathy J. Price. A review and synthesis of the first 20years of PET and fMRI studies of heard speech, spoken language and reading. NeuroImage, 62 0 (2): 0 816--847, August 2012. ISSN 1053-8119. doi:10.1016/j.neuroimage.2012.04.062. https://www.sciencedirect.com/science/article/...

    2012 doi

  40. [48]

    Child, D

    Alec Radford, Jeff Wu, R. Child, D. Luan, Dario Amodei, and I. Sutskever. Language Models are Unsupervised Multitask Learners . OpenAI blog, 2019

    2019

  41. [49]

    The WEAVER model of word-form encoding in speech production

    Ardi Roelofs. The WEAVER model of word-form encoding in speech production. Cognition, 64 0 (3): 0 249--284, September 1997. ISSN 00100277. doi:10.1016/S0010-0277(97)00027-9. https://linkinghub.elsevier.com/retrieve/pii/S0010027797000279

    1997 doi

  42. [50]

    Rumelhart and Donald A

    David E. Rumelhart and Donald A. Norman. Simulating a Skilled Typist : A Study of Skilled Cognitive - Motor Performance . Cognitive Science, 6 0 (1): 0 1--36, 1982. ISSN 1551-6709. doi:10.1207/s15516709cog0601_1. https://onlinelibrary.wiley.com/doi/abs/10.1207/s15516709cog0601...

    1982 doi

  43. [51]

    Russo, Ramin Khajeh, Sean R

    Abigail A. Russo, Ramin Khajeh, Sean R. Bittner, Sean M. Perkins, John P. Cunningham, L. F. Abbott, and Mark M. Churchland. Neural Trajectories in the Supplementary Motor Area and Motor Cortex Exhibit Distinct Geometries , Compatible with Different Classes of Computation . Neu...

    2020 doi

  44. [52]

    Dynamics of brain activation during picture naming

    Riitta Salmelin, R Hari, O V Lounasmaa, and Mikko Sams. Dynamics of brain activation during picture naming. Nature, 368: 0 463--465, 1994

    1994

  45. [53]

    Attention is All you Need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. Attention is All you Need . In Advances in Neural Information Processing Systems , volume 30. Curran Associates, Inc., 2017. https://proceedings.neurip...

    2017

  46. [54]

    Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, Stéfan J

    Pauli Virtanen, Ralf Gommers, Travis E. Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, Stéfan J. Van Der Walt, Matthew Brett, Joshua Wilson, K. Jarrod Millman, Nikolay Mayorov, Andrew R. J. Nelson, E...

    2020

  47. [55]

    McMahon, and Greig I

    Angelique Volfart, Katie L. McMahon, and Greig I. De Zubicaray. A Comparison of Denoising Approaches for Spoken Word Production Related Artefacts in Continuous Multiband fMRI Data . Neurobiology of Language, 5 0 (4): 0 901--921, September 2024. ISSN 2641-4368. doi:10.1162/nol_...

    2024 doi

  48. [56]

    Removal of Muscle Artifacts from EEG Recordings of Spoken Language Production

    De Maarten Vos, Stephanie Riès, Katrien Vanderperren, Bart Vanrumste, Francois-Xavier Alario, Van Sabine Huffel, and Boris Burle. Removal of Muscle Artifacts from EEG Recordings of Spoken Language Production . Neuroinformatics, 8 0 (2): 0 135--150, June 2010. ISSN 1539-2791, 1...

    2010 doi

  49. [57]

    Wikimedia Downloads

    Wikidumps. Wikimedia Downloads . https://dumps.wikimedia.org/

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.