Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 89 of the 89 references it cites were found in a registry or an index.

Hierarchical Organization of Auditory and Motor Representations in Speech Perception: Evidence from Searchlight Similarity Analysis

Open access
Evans, Samuel, Davis, Matthew H. · 2015 · Cerebral Cortex

Abstract

How humans extract the identity of speech sounds from highly variable acoustic signals remains unclear. Here, we use searchlight representational similarity analysis (RSA) to localize and characterize neural representations of syllables at different levels of the hierarchically organized temporo-frontal pathways for speech perception. We asked participants to listen to spoken syllables that differed considerably in their surface acoustic form by changing speaker and degrading surface acoustics using noise-vocoding and sine wave synthesis while we recorded neural responses with functional magnetic resonance imaging. We found evidence for a graded hierarchy of abstraction across the brain. At the peak of the hierarchy, neural representations in somatomotor cortex encoded syllable identity but not surface acoustic form, at the base of the hierarchy, primary auditory cortex showed the reverse. In contrast, bilateral temporal cortex exhibited an intermediate response, encoding both syllable identity and the surface acoustic form of speech. Regions of somatomotor cortex associated with encoding syllable identity in perception were also engaged when producing the same syllables in a separate session. These findings are consistent with a hierarchical account of how variable acoustic signals are transformed into abstract representations of the identity of speech sounds.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/cercor/bhv136

Cites

89 of 89 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. (A) Spectrograms of /ba/ syllables presented as clear speech (CL), NV speech (NV), and SW resynthesized (SW). (B) Behavioral data i) one-back repetition detection accuracy during scanning and ii) proportion correct identification of degraded syllables after scanning (horizontal line shows chance level of 0.16). (C) Block structure showing trial sequences during an example scanning run. Red = NV syllables, blue = SW syllables, green = CL syllables, grey = null trials. Dark coloring = male speaker, light coloring = female speaker. Grey outer shading plus button icon indicates repetition trials. (D) Single-trial structure showing the timing of stimulus presentation and MR volume acquisition. (E) Experiment structure showing odd- and even-numbered experimental blocks in each scanning run.
Figure 1. (A) Spectrograms of /ba/ syllables presented as clear speech (CL), NV speech (NV), and SW resynthesized (SW). (B) Behavioral data i) one-back repetition detection accuracy during scanning and ii) proportion correct identification of degraded syllables after scanning (horizontal line shows chance level of 0.16). (C) Block structure showing trial sequences during an example scanning run. Red = NV syllables, blue = SW syllables, green = CL syllables, grey = null trials. Dark coloring = male speaker, light coloring = female speaker. Grey outer shading plus button icon indicates repetition trials. (D) Single-trial structure showing the timing of stimulus presentation and MR volume acquisition. (E) Experiment structure showing odd- and even-numbered experimental blocks in each scanning run.CC BY · PubMed Central
Figure 2. (A) Speech perception network: [Syllable listening > Null events] rendered onto a canonical brain image. (B) Speech production network: [Overt > Covert]. (C) Effect of degradation: [((NV + SW)/2) > CL]. Response plot shows parameter estimates from the peak voxel in the LIFG and premotor cluster with error bars suitable for repeated-measures comparisons (Loftus and Masson 1994), and a scatter plot of the correlation between the BOLD response from the peak voxel and the d′ scores for the identification of repetitions of degraded syllables. (D) Whole-brain correlation relating d′ scores for degraded syllables and brain activity for degraded speech, with a scatter plot of the data within the Inferior Parietal Lobe region showing a significant correlation (for purpose of illustration not for effect size inference). All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.
Figure 2. (A) Speech perception network: [Syllable listening > Null events] rendered onto a canonical brain image. (B) Speech production network: [Overt > Covert]. (C) Effect of degradation: [((NV + SW)/2) > CL]. Response plot shows parameter estimates from the peak voxel in the LIFG and premotor cluster with error bars suitable for repeated-measures comparisons (Loftus and Masson 1994), and a scatter plot of the correlation between the BOLD response from the peak voxel and the d′ scores for the identification of repetitions of degraded syllables. (D) Whole-brain correlation relating d′ scores for degraded syllables and brain activity for degraded speech, with a scatter plot of the data within the Inferior Parietal Lobe region showing a significant correlation (for purpose of illustration not for effect size inference). All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.CC BY · PubMed Central
Figure 3. Similarity of multivariate patterns for the acoustic form (left) and syllable identity (right) representational similarity analysis models. RSA models express expected differences in voxel pattern similarity between neural responses to different syllables presented in different acoustic forms in odd and even blocks of trials. Red tiles indicating increased similarity between pairs of neural patterns, blue tiles indicating decreased similarity, green tiles indicating comparisons that were removed from analysis as not applicable (N/A). Results of searchlight RSA for these 2 models are shown rendered onto a canonical brain image. All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.
Figure 3. Similarity of multivariate patterns for the acoustic form (left) and syllable identity (right) representational similarity analysis models. RSA models express expected differences in voxel pattern similarity between neural responses to different syllables presented in different acoustic forms in odd and even blocks of trials. Red tiles indicating increased similarity between pairs of neural patterns, blue tiles indicating decreased similarity, green tiles indicating comparisons that were removed from analysis as not applicable (N/A). Results of searchlight RSA for these 2 models are shown rendered onto a canonical brain image. All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.CC BY · PubMed Central
Figure 4. (A) Univariate activity associated with speech production: [Overt > Covert] (green), signal degradation during speech perception: [((NV + SW)/2) > CL] (Red), and multivariate syllable identity coding (blue) displayed on a canonical brain image. Note that multivariate decoding maps were resampled to 2-mm isotropic voxels to illustrate and quantify voxel-to-voxel overlap between univariate and multivariate results. i) Overlap between neural responses to signal degradation and speech production, ii) overlap between neural coding of syllable identity and neural responses during speech production and iii) coronal and sagittal sections showing the adjacent location of activation of responses to speech degradation during perception and syllable identity coding. (B) Comparison of the similarity response in the bilateral PAC and the left somatomotor cortex for the syllable identity and acoustic form models. Bar graphs show similarity response in these regions with error bars showing within-subject standard error (Loftus and Masson 1994). All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.
Figure 4. (A) Univariate activity associated with speech production: [Overt > Covert] (green), signal degradation during speech perception: [((NV + SW)/2) > CL] (Red), and multivariate syllable identity coding (blue) displayed on a canonical brain image. Note that multivariate decoding maps were resampled to 2-mm isotropic voxels to illustrate and quantify voxel-to-voxel overlap between univariate and multivariate results. i) Overlap between neural responses to signal degradation and speech production, ii) overlap between neural coding of syllable identity and neural responses during speech production and iii) coronal and sagittal sections showing the adjacent location of activation of responses to speech degradation during perception and syllable identity coding. (B) Comparison of the similarity response in the bilateral PAC and the left somatomotor cortex for the syllable identity and acoustic form models. Bar graphs show similarity response in these regions with error bars showing within-subject standard error (Loftus and Masson 1994). All results presented at P < 0.001 peak level uncorrected, q < 0.05 cluster-level FDR corrected.CC BY · PubMed Central
Figure 5. Hierarchical abstraction of syllable encoding in the left precentral gyrus and the left and right temporal cortex. (A) Brain rendering showing the overlap between the leave-one-participant-out ROIs used in the analysis. These regions showed a significant fit to the full syllable identity coding model (Fig. 3, right panel) but are defined for each participant using data from all other participants. This ensured that follow-up tests comparing different submodels are statistically independent of the data used to define the ROI. We differentiated the response in these regions by testing for 4 component models shown in (B). These test for similarity driven for identical syllables presented in the Same Degradation-Same Speaker (Model 1), Different Degradation-Same Speaker (Model 2), Same Degradation-Different Speaker (Model 3), and Different Degradation-Different Speaker (Model 4, the most abstract form of similarity tested in the current design). Red tiles indicate cells in which increased similarity is expected, blue tiles indicate decreased similarity, and green tiles are comparisons removed from analysis. Bar graphs show similarity responses for each of these 4 models in each ROI with within-subject error bars representing one standard error of the mean suitable for repeated-measures comparisons (Loftus and Masson 1994).
Figure 5. Hierarchical abstraction of syllable encoding in the left precentral gyrus and the left and right temporal cortex. (A) Brain rendering showing the overlap between the leave-one-participant-out ROIs used in the analysis. These regions showed a significant fit to the full syllable identity coding model (Fig. 3, right panel) but are defined for each participant using data from all other participants. This ensured that follow-up tests comparing different submodels are statistically independent of the data used to define the ROI. We differentiated the response in these regions by testing for 4 component models shown in (B). These test for similarity driven for identical syllables presented in the Same Degradation-Same Speaker (Model 1), Different Degradation-Same Speaker (Model 2), Same Degradation-Different Speaker (Model 3), and Different Degradation-Different Speaker (Model 4, the most abstract form of similarity tested in the current design). Red tiles indicate cells in which increased similarity is expected, blue tiles indicate decreased similarity, and green tiles are comparisons removed from analysis. Bar graphs show similarity responses for each of these 4 models in each ROI with within-subject error bars representing one standard error of the mean suitable for repeated-measures comparisons (Loftus and Masson 1994).CC BY · PubMed Central
Figure 6. Structure of syllable identity representations within the left precentral gyrus and the left and right temporal cortex. (A) Brain rendering showing the overlap between the leave-one-participant-out ROIs used in the analysis. We further differentiated the response in regions shown to encode syllable identity to identify the additional structure of syllable identity representations (B) within these regions. These test for similarity driven by the place of articulation feature (Model 1) and manner of articulation feature for consonants (Model 2), phoneme identity (Model 3) and CV vs. VC syllable structure (Model 4). Red tiles indicate cells in which increased similarity is expected, blue tiles indicate decreased similarity, and green tiles are comparisons removed from analysis. Bar graphs show similarity responses for each of these 4 models in each ROI as plotted in Figure 5.
Figure 6. Structure of syllable identity representations within the left precentral gyrus and the left and right temporal cortex. (A) Brain rendering showing the overlap between the leave-one-participant-out ROIs used in the analysis. We further differentiated the response in regions shown to encode syllable identity to identify the additional structure of syllable identity representations (B) within these regions. These test for similarity driven by the place of articulation feature (Model 1) and manner of articulation feature for consonants (Model 2), phoneme identity (Model 3) and CV vs. VC syllable structure (Model 4). Red tiles indicate cells in which increased similarity is expected, blue tiles indicate decreased similarity, and green tiles are comparisons removed from analysis. Bar graphs show similarity responses for each of these 4 models in each ROI as plotted in Figure 5.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 89 references

Research works printed but not linked: 2

  1. 10.1037/0096-1523.23.3.914 · doi:10.1037/0096-1523.23.3.914 · Paper page
  2. Zevin, 2004, Dishabituation to phonetic stimuli in a “silent” event-related fMRI design, Int J Psychol, 39, 102

Cited by

See all 133

Metadata

Authors
Evans, S., & Davis, M. H.
Affiliations
MRC Cognition and Brain Sciences Unit (GB)
Corresponding author
Evans, Samuel
Venue
Cerebral Cortex
Published
2015-07-08
Volume
25
Issue
12
Pages
4772-4788
ISSN
  • 1047-3211 (ISSN-L)
  • 1460-2199
OpenAlex
W2197518923
Cited by
130
References
89
Citations by year
2026: 6 · 2025: 12 · 2024: 5 · 2023: 12 · 2022: 18 · 2021: 9 · 2020: 10 · 2019: 12 · 2018: 11 · 2017: 19 · 2016: 13 · 2015: 3
FWCI
9.22
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Cerebral Cortex (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • WestminsterResearch (University of Westminster) (acceptedVersion, cc-by)
  • UCL Discovery (University College London) (submittedVersion)
Funders
Medical Research Council (MC-A060-5PQ80)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neurobiology of Language and Bilingualism
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Phonetics and Phonology Research
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neuroscience and Music Perception
Keywords
speech perception, hierarchical representation, somatomotor cortex, primary auditory cortex, temporal cortex, noise-vocoded speech, sine-wave speech, speech production, representational similarity analysis
MeSH terms
Adolescent, Adult, Auditory Threshold, Data Interpretation, Statistical, Female, Frontal Lobe, Humans, Male, Neural Pathways, Noise, Speech, Speech Perception, Temporal Lobe, Multivariate Analysis, Young Adult, Sensorimotor Cortex

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Hierarchical Organization of Auditory and Motor Representations in Speech Perception: Evidence from Searchlight Similarity Analysis. https://app.cochat.ai/papers/10.1093/cercor/bhv136
The paper
Evans, S., & Davis, M. H. (2015). Hierarchical Organization of Auditory and Motor Representations in Speech Perception: Evidence from Searchlight Similarity Analysis. Cerebral Cortex. https://doi.org/10.1093/cercor/bhv136

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.