Work Description

Title: Cortical 𝛾-Oscillations Implement Basic Language Operations: Evidence from Electroencephalography in Anaphora During English Filler-gap Dependency Processing Open Access Deposited

http://creativecommons.org/licenses/by-nc/4.0/
Attribute Value
Abstract
  • Background: Top-down structure-building hypotheses propose that gamma (>30Hz) oscillations in frontotemporal circuitry create syntax-semantics objects, contrasting with slow-rhythm entrainment of gamma oscillations tracking semantic fitness in bottom-up chunking. In bi-clausal filler-gap dependencies, like which message regarding/about him did Frank/Diana say__ that Diana/Frank received__unexpectedly? wh-fillers, like which message regarding/about him, link to gap__ sites in re-representation. The pronoun him, as a contextually interpreted expression, receives an interpretation in syntax or discourse. Crucially, adjunct noun-phrase-modifiers (Mods) like regarding him and lexically specified noun-complements (Comps) like about him have equal contextual fitness. They are, however, biased toward discourse-coreference (Mods) vs. syntax-binding (Comps). In coreference, discourse referents for each expression map to the same entity. In syntactic binding, multiple expressions map to a single entity in one fell swoop. Binding him to Frank requires forming a referential combination {Frank, him} mapping to a discourse referent x for an entity named Frank, creating the referential object {x, {Frank, him}}. Anaphora resolution has been tied to increased power at 30-50Hz in retrieval of referential elements and 50-90Hz in integration. This predicts gamma-range Event-Related Power Differences (ERPDs) in filler-gap processing as anaphora processes for Mod/Comp interact with early/late-antecedent match. Methods: EEG data were recorded at a 1000Hz sampling rate via a 64-electrode system from 23 right-handed English speakers. Using cluster-based nonparametric permutation tests, Mod-Comp based anaphora-linked ERPDs were examined at 30-50Hz and 50-90Hz in evoked activity at the gap sites say__ that and received__ to capture cell assembly formation in referential object creation for Mods vs. Comps in wh-filler re-representations. Results: Gamma-power differences for Mods vs. Comps in early vs. late antecedent match arose 30-50Hz (p = 0.013, t = -469.98, duration 70ms) and 50-70Hz (p = 0.042, t = -198.79, duration 35ms) at say__ that upon accessing say, and 50-90Hz (p = 0.007, t = -411.3, duration 75ms) at received__ upon accessing received in final integration. Conclusions: Despite equal semantic fitness for Mods and Comps, Mod-Comp modulations of ERPDs linked to early vs. late antecedents arose in referential processing aligned with gap positions, suggesting the gamma-based implementation of referential relations in complex dependencies.
Methodology
  • 4.1. Participants and testing procedures A total of 27 participants with English as their first language were recruited for this study. One participant aborted the study, leaving the total participant count to 26 (14 F 12 M; 23 RH 3 LH). We report results from 23 right-handed participants. These participants were undergraduate students with no history of dyslexia. Accuracy rates of 76% on factual comprehension checks show the task to be challenging. On comprehension checks related to anaphoric interpretation, respondents interpreted the pronoun as referring to the gender-matched noun phrase 75% of the time. However, comprehension check accuracy was not used as a filter for analysis: Real-time brain processing as participants compute a bi-clausal filler-gap dependency is expected to be independent of their behavior on comprehension checks following individual sentences. 4.2. EEG preprocessing and data extraction EEG was recorded at a 1000Hz sampling rate via a 64-electrode EGI system (Electrical Geodesics Inc., Eugene, OR; as displayed in Fig. 1) referenced to Cz (vertex) online. The signal was collected using a Net Amps 300 amplifier with a gain of 5000 and acquisition software Netstation (version 4.5.4). Impedances were verified to be below 50 kΩ before each of the four blocks in the task. All preprocessing and data cleaning procedures were performed using the EEGLAB toolbox based on MATLAB (version 9.5). An 8ms latency shift due to the amplifier was corrected before preprocessing. Line noise was removed using the CleanLine plugin for EEGLAB. The continuous data were then divided into 3.850-second epochs starting with did (the question marker) and running to the end of the sentence. These epochs included the time windows of two words of interest: the bridge verb say and the embedded-clause thematic verb received, including a baseline period of two critical words did and the matrix-clause subject. Following segmentation, we visually inspected each epoch for bad channels and, if a channel was bad in more than 10% of epochs, we removed the whole channel. It has been demonstrated that muscle activity can create noise in high-frequency EEG measures, and γ-band results should thus be reported and interpreted with caution. Hipp and Siegel showed that removing such artifacts from the EEG recording through rejecting data sections affected by artifactual signals or Independent Component Analysis (ICA) can allow for more confident analysis of high-frequency EEG. Therefore, we visually inspected each epoch and systematically removed any epoch including unexpected EMG activity (i.e., furrowing of the brow, face and neck movements, but not blinks). We used the binica algorithm for ICA to extract and then manually check the 32 most impactful components generated by principal component analysis (PCA) so that we could effectively remove the remaining ocular and cardiac activity, among any remaining artifacts. An average of 89% of trials was retained across subjects (SD = 3.120). The average number of trials retained was similar across conditions, 1a, M = 27.00, SD = 3.37; 1b, M = 26.61, SD = 3.60; 1c, M = 25.83, SD = 3.08; 1d, M = 27.26, SD = 2.26. The average rejection number per subject was 19.01. The data were average referenced and missing channels were interpolated for the time-frequency analysis. <INSERT FIG. 1 ABOUT HERE> Fig 1. The 64-electrode EGI system. 4.3. Time-frequency analysis The preprocessed EEG data were loaded into the FieldTrip toolbox as four datasets for the four structural conditions (2a - d). The time-frequency analysis was run in the time window of 3,850ms starting from 270ms into the word ‘did’ to the end of the sentence to achieve stable frequency decomposition. We convolved a family of Morlet wavelets of 7 cycles in .5 Hz steps within the selected time window of each EEG trial, which yielded the time-frequency information of the neural activity. The length of the wavelets was set as 3 standard deviations of the Gaussian kernel. At 60 Hz, the wavelet duration was 0.037 seconds. The spectral bandwidth was 17.143 Hz. We log-transformed (10*log10) the derived power in Fieldtrip to standardize the unit as decibels at each of the frequencies between 30 Hz and 120 Hz for each condition of each subject. We extracted the power information for 500ms starting at 200ms into the presentation of each verb, marking the start of lexical access and emergence of processes in abstract syntax, conceptual structure, and semantic interpretation. These time windows capture processing moments where the bridge verb enables an embedded clause dependency through an intermediate gap position, and when the embedded clause verb enables a thematic gap position enabling the thematic integration of the wh-filler. Following Dekydtspotter et al., we examined evoked γ power that reflects both time-locked and phase-locked oscillatory responses. We calculated the evoked power of each condition for each subject by removing the induced power from the total power. 4.4. Data analysis Data were analyzed with cluster-based nonparametric permutation tests to avoid the multiple comparison problem for our medium-density electrodes, on the assumption that the spatially adjacent channels exhibit similar spectral-temporal features. We conducted paired-samples t-tests using Monte Carlo simulations with 1000 random samplings for each channel-frequency-time triplet. As we are interested in the high/broadband hierarchical processing across the γ band and the narrowband processes of retrieval in low γ frequencies, we used bins for low γ 30-50 Hz and high γ 50-90 Hz. Therefore, we calculated Mod-Comp power differences between the antecedent match [(2a) - (2b)] and antecedent mismatch [(2c) - (2d)] conditions to address distinct allocations of resources to Mods vs. Comps. We then examined differences between Mod and Comp structures in early vs. late antecedent match using paired-sample t-tests using the maximum of the cluster t-test statistics with 1000 permutations. We adopted a Bonferroni correction of α = .025 for two bins. Permutation tests provide the time window and electrodes in which a significant effect arises. They, however, lack precision as to the exact timing and location of effects. Our discussion of the timing of ERPD effects is therefore limited to the window in which the effects are found.
Description
  • This study addresses a debate on the nature of gamma (>30Hz) oscillations in language generation and structuring from a top-down vs. bottom-up chunking approach. Working within the framework of recent oscillatory models for syntax, we argue that a time-frequency analysis of electroencephalography at crucial points of recursions in bi-clausal wh-movement is probative of the role of gramma oscillations in the creation of referential dependencies nested into wh-dependencies, accounting for aspects of the expressive power of language. The study proposes that gamma oscillations in frontotemporal circuitry create syntax-semantics objects in contrast with slow-rhythm entrainment of gamma oscillations tracking semantic fitness in bottom-up chunking. We show gamma power effects associated with aspects of anaphoric relations established in syntax vs. discourse during the processing of bi-clausal filler-gap dependencies. Despite equal semantic fitness across conditions for noun phrase modifiers (Mods) and noun complements (Comps), Mod-Comp modulations of ERPDs linked to anaphoric relations with early vs. late antecedents arose in referential processing aligned with gap positions, suggesting the gamma-based implementation of referential relations is induced by the re-representations of pronouns in complex filler-gap dependencies. 
Creator
Depositor
Contact information
Keyword
Citations to related material
Publisher
Resource type
Last modified
  • 09/05/2024
License
To Cite this Work:
Dekydtspotter, L., Miller, A., Meinert, K., Cha, J., Ah, J., Swanson, K., Xiong. Yanyu Cortical 𝛾-Oscillations Implement Basic Language Operations: Evidence from Electroencephalography in Anaphora During English Filler-gap Dependency Processing [Data set]. Indiana University - DataCORE.

Relationships

Files (Count: 4; Size: 4.31 GB)

Thumbnail Title Original Upload Last Modified File Size Access Actions
Comp_match_b.mat 2024-09-05 2024-09-05 1.08 GB Open Access
Comp_mismatch_d.mat 2024-09-05 2024-09-05 1.08 GB Open Access
Mod_match_a.mat 2024-09-05 2024-09-05 1.08 GB Open Access
Mod_mismatch_c.mat 2024-09-05 2024-09-05 1.08 GB Open Access

Download All Files

All files in this dataset will be included in the zip download.
Files can be downloaded individually in the "Files" panel above.

Total work file size of 4.31 GB may be too large to download directly. Consider using Globus (see below).